Absenteeism amongst health workers – developing a typology to support empiric work in low-income countries and characterizing reported associations

The contribution of inadequate health worker numbers and emigration have been highlighted in the international literature, but relatively little attention has been paid to absenteeism as a factor that undermines health-care delivery in low income countries. We therefore aimed to review the literature on absenteeism from a health system manager’s perspective to inform needed work on this topic. Specifically, we aimed to develop a typology of definitions that might be useful to classify different forms of absenteeism and identify factors associated with absenteeism. Sixty-nine studies were reviewed, only four were from sub-Saharan Africa where the human resources for health crisis is most acute. Forms of absenteeism studied and methods used vary widely. No previous attempt to develop an overarching approach to classifying forms of absenteeism was identified. A typology based on key characteristics is proposed to fill this gap and considers absenteeism as defined by two key attributes, whether it is: planned/unplanned, and voluntary/involuntary. Factors reported to influence rates of absenteeism may be broadly classified into three thematic categories: workplace and content, personal and organizational and cultural factors. The literature presents an inconsistent picture of the effects of specific factors within these themes perhaps related to true contextual differences or inconsistent definitions of absenteeism.

To assess the consequences of work-related skin and airway symptoms among these dentists, in terms of contact with health authorities, sick leave, or changes in professional career. 3082 of 3500 dentists in public and private practice Questionnaires sent to 3500 dentists. Phone interviews with 19 dentists who reported change of career due to work related skin and airways symptoms.
MSD known to cause sickness absence among dentists. Skin and airways infections did not significantly contribute to sickness absence and change of career.

Albion MJ 2008, Australia
To explore relations among positive and negative psychological states and both absenteeism and turnover intentions.
1097 of 1683 hospital employees were surveyed using the Queensland Public Agency Staff Survey (QPASS) Survey used to obtain measures of organizational climate, psychological reactions to work, job satisfaction, and selfreported levels of intention to leave.
The hypothesised mediation effect of psychological states on the relationship between organizational climate and absenteeism did not emerge.

Alexopoulos 2011, Greece
To investigate the relationships between physical, psychosocial, and individual characteristics and occurrence, chronicity, care seeking and absenteeism due to musculoskeletal complaints of the lower back and knee.

of 448 nursing personnel in six hospitals
A self administered questionnaire was distributed by the researchers between September and December 2007.
Perceived moderate/bad general health and high need for recovery were risk factors for absenteeism due to low back pain. Organizational factors, high job demands and low co workers support were risk factors for absenteeism due to knee pain Al-Shammari, Bamgboye et al. 1994, Saudi Arabia To analyse the sickness absence records of medical staff in a university teaching hospital and identify the categories of hospital personnel that are at high risks and causes for sickness absenteeism.
Workers in a hospital (contract and noncontract) including admin and support staff. N=861 The sickness absence records of workers at the King Khalid University Hospital over a period of two years (1990)(1991) were analysed to identify the category of workers at high risks.
Prevalence of sickness absence higher among contract workers than noncontract workers. The sex ratio was 1:2 in favour of females among contract employees but 1.2:1 in the male's favour for the Saudis. Females have more spells of sickness than males and the duration of sickness absence is significantly higher in females. No significant occupational effect on the spells and duration of sickness absence. Anagnostopoulos 2010, Greece To examine how job burnout was associated with sick-leave Nurses n=487 Cross-sectional study. Use of questionnaire survey measuring 54% had no sickness absence recorded in the last 12 months, 37.6% had taken absence and quality of life burnout levels and quality of life. These were then matched to absence data from hospital records up to 10 days, and 8.4% had taken more than 10 days. Significant negative correlations observed between sickness absence duration and mental and physical health Nurses with high burnout levels were considerably more likely to be absent from work for a short period than those with low burnout levels Bourbonnais 1992, Canada To test the hypothesis:-That among nurses, the heavier the work load, the more sick leave nurses would take Nurses from 7 general hospitals in Quebec who had been employed for at least 6 months.
Case-control study Cases (experienced at least one episode of medically certified sick leave for a diagnosis "most likely to be related to work load" between 1. 01.1984 and 31.05.1987) n=184, Controls n=1165 Significant associations found between sick leave and nurse patient ratio among head nurses, patient's duration of stay, and one hospital. Sick leave more frequent among full time permanent nurses and among those on night and evening schedules Head nurses more absent than nurses Bourbonnais 2001, Canada To determine whether nurses exposed to job strain had a higher incidence of sick leave than nurses not exposed To determine if social support at work modifies the association between job strain and the incidence of sick leave among nurses Nurses n=1891 of 3065 1793 of 1891 consented to examination of absence records Longitudinal study. Use of questionnaires and examining absence records Short-term sick leaves associated with job strain and with low social support at work. Certified sick leaves also significantly associated with low social support at work Social support at work does not modify the association between job strain and absence Bamgboye and Adeleye 1992, Nigeria To fill the gap in the knowledge of sickness absence records and statistics among Nigerian hospital workers.
Employees in a University Teaching Hospital n=3112.
The sickness absence records of employees in a University Teaching Hospital in Nigeria were examined over a period of three years.
An overall proportion of absentee workers was 15.8% with an average of 3 spells of sickness per year per absentee while the duration of sickness per absentee was 5.6 days per year. The younger employees less than 35 years of age and those with short duration of employment with the hospital have significantly higher spells and duration of sickness absence than others. Lower spell of sickness and duration of sickness absence were observed among nurses, senior employees especially doctors had no records of sickness absenteeism in any of the 3 years of study. Bamgboye, Olubuyide et al. 1993 There was a sickness absence rate of 7.7% with an average of 0.24 spells of sickness and duration of 1.24 days per absentee per year. The biological characteristics of the absentees showed no differentials in the spell rate of sickness, but nurses and junior workers had higher duration of sickness absence than other categories of workers. Acute respiratory and upper tract infections, sore throat, chickenpox, septicaemia, and diseases of the musculoskeletal system (mainly back pain) were the leading conditions resulting in absenteeism. Becker and Oliveira 2008, Brazil To research the rate of absenteeism of nursing professionals in a psychiatric center in Manaus, from January/2004 to January/2005, in the Human Resources sector of the institution

Nurses n=105
The records of workers who were absent from their professional activities for at least one day were checked in the study period Sickness absence more common in respondents who reported health complaints at baseline, smokers, those with low personal engagement in the work unit, lack of encouraging and supportive culture in the work unit, perceived unsuspicious culture in the work unit, working in psychiatric and paediatric wards, having injured the neck in an accident, having changed work or work tasks because of pain, musculoskeletal pain, and fatigue Responders who engaged in aerobics or gym less likely to be absent from work. Franche 2011, Canada To evaluate the impact of worker and workplace factors and of their relationships on work absence duration Female nurses engaged in direct care n=11762 of 15167 Structural equation modelling (a telephone survey completed) Worker health factors, namely pain related work interference, work-related pain, and depression, had the largest total effects on work absence duration. Emotional abuse or physical assault from a patient or visitor had the largest total effect on work absence duration of all workplace factors Low respect and support at work and abuse or assault by a co-worker.

Fujishiro 2011, Philippines
To examined whether the experience of workplace aggression was associated with poor self-rated health, as well as with work-related injury or illness, and whether being exposed to physical assault and verbal aggression simultaneously was associated with a higher risk of workrelated injury or illness, compared with being exposed to either type of aggression alone.
Nurses attending the 2007 national convention of the Philippine Nurses Association n=690 of 1000 Self administered questionnaire. Cross sectional design as the nurses were from 13 regions of the Philippines Those who reported experiencing physical assault were more likely to have missed more than 2 workdays in the past year. Targets of verbal abuse showed a trend toward missing work

Garcia-Prado and Chawla 2006, Costa Rica
To test whether the introduction of these new organizational arrangements was followed by lower absence rates than those registered previously in Costa Rica Physicians and healthcare workers in 29 public hospitals.
Data on health care personnel absence, specifically cumulative sick-leave days were obtained. Average absence figures were then computed for each hospital and year by dividing total days of absence by total number of health staff.
Reforms had a negative impact on absenteeism, which increased throughout the considered period. The policy of not substituting absentee workers, which was introduced through the reforms, did not work as expected in a permissive environment in which peer pressure mechanisms were lacking.

Gaudine and Gregory 2010, Canada
To determine the accuracy of nurses' self-reports of absence

Nurses n= 215
The study design is correlational and descriptive. Self-reported absence from questionnaires completed nurses then compared with organizational records.
Strong positive correlation, a strong intra-class correlation and Cronbach's alpha for the two measures of absence. Difference in central tendency that is related to the majority of nurses in this study underestimating their days absent from work. Gorman, Yu et al.

2010, British Columbia, Canada
To determine the demographic and work characteristics of healthcare workers who were more likely to take sickness absences from work.
Healthcare workers (n= 36,858). In 3 health authorities. 59.2% full-time and 40.8% part-time employees Cross-sectional study design. Payroll data were analyzed for three health regions. Sickness absence rates were determined per person-year and then compared across demographic and work characteristics using multivariate Poisson regression models. The direct costs to the employer due to sickness absences were also estimated.
Female, older, full-time workers, longterm care workers and those with a lower hourly wage were more likely to take sickness absences and had similar trends with respect to the costs due to sickness absence. For occupations, licensed practical nurses, care aides and facility support workers had higher rates of sickness absence. Registered nurses, and those workers paid high hourly wages were associated with highest sickness related costs. Genevay 2011, Switzerland To explore the staff-and work-related risk factors for spinal pain among hospital employees, To investigate the effect of staffand work-related variables on the consequences of spinal pain, such as doctor visits and sick leave To ascertain whether the same individual factors, working conditions and health problems had led to increased probability of both leaving jobs and prolonged sickness absence in a cohort of Swedish nurses over a period of 3 years.

Nurses and assistant nurses (n=2293)
A baseline questionnaire was answered by 2293 nurses, representing a response rate of 86%. Exposed and unexposed nurses were compared with regard to two outcomes. During the 3-year follow-up, exposed and unexposed nurses were compared with regard to two outcomes: resigning and having at least one sick leave spell that lasted 28 days or longer.
18% of the nurses left their employment, and 16% had sick leave spells ≥28 days. Work in geriatric care, being socially excluded by superiors and/or workmates, negative effects of organizational changes and poor selfrated general health were factors that increased the likelihood of both leaving jobs and long-term sick leave. During the study period (1981)8146 reported spells of absence with 34 829 days lost were attributed to sickness. A comparison with the data of the study performed in the same hospital about 15 years ago showed a rise in the duration of absences with a simultaneous reduction in their incidence. Sickness absenteeism was higher among female, mainly unskilled, workers, presently or previously married, aged from 45 to 60, and employed in the hospital for over ten years.

Rajbhandary 2010, Canada
To investigates the relationship between the working conditions and illness-and injury-related Registered and licensed practical and registered psychiatric nurses n= Data extracted from a national survey of the work and health of nurses i.e including absence data The average absenteeism for RNs is less than those for LPNs. Role overload score and effort-reward absenteeism among nurses 18,676 of 23428 After restricting the sample to fulltime employed RNs and LPNs and eliminating those with missing data total participants n=6634 imbalance score significant for LPNs but not for RNs. The province in which one works influences their absence rates. Those working in Newfoundland and Those working in hospitals have higher absence rates. Females more absent than males. Rauhala, Kivimäki et al. 2007, Finland To examine whether nurses' work overload is associated with increased sick leave and quantifying the loss of working days from work overload.

Nurses (registered and practical) n=877
An observational cohort study Patient-associated workload scores from the RAFAELA system were based on a 6-month monitoring period in 2004. Records of 12-month sick leave in the same year were obtained from employers' registers.
The mean workload was 9% above the optimum. There was a linear trend between increasing workload and increasing sick leave. Among nurses with workload ≥30% above the optimum the rate of self certified periods of sick leave was 1.44 times higher than among those with an optimum workload. The corresponding rate ratio for medically certified sick leave was 1.49 These excess rates of sickness absence resulted in 12 extra sick leave days per person-year. Remsburg, Armacost et al.

1999, USA
To identify reasons for absenteeism among nursing assistants.
To identify the impact that absenteeism had on achieving daily target hours of care. To identify strategies to reduce absenteeism.
Nurse Assistants in Longterm care units The following data were collected: (1) number of daily call-ins, (2) reason for call-in, and (3) the percentage of shifts not meeting daily targeted hours of care.
A total of 300 call-ins during the 3month period. Only 6 days during the 3month period had no call-ins. Number of daily call-ins ranged from 0 to 9, with a mean of 3.3 call-ins per day. Approximately 79 nursing assistants were assigned for duty each weekday; therefore, the average rate of absenteeism was 4%. Ritchie, Macdonald et al. 1999, UK To identify the healthcare needs of healthcare workers using routinely collected sickness absence data Nurses, ancillary, administrative and clerical staff, and medical and dental staff (n=12900) Analysis of sickness absence records for one year -1993 Female staff more likely to have experienced absence than male staff. Full time staff had greater rates of sickness absence than part time staff. Medical and dental staff had the lowest rate of recorded sickness absence.
Maintenance staff had proportionately the greatest number of absences of >1 week and pharmacy staff had the fewest of these longer term absences. The main known causes of absence were respiratory disorders, digestive disorders, and musculoskeletal disorders.

Schalk 2011, Netherlands
To examine the influence over time of organizational commitment, health complaints, and visits to a general practitioner on sickness absenteeism.

Nurses n= 224
A longitudinal, three-wave study in two nursing homes. Questionnaire data (self reports of organizational commitment, health complaints, visits to a general practitioner) Absenteeism data retrieved from personnel files.
Health complaints and visits to a general practitioner were found to predict absenteeism behavior. Commitment was related to health complaints at the same point in time, but did not predict future sickness absenteeism. Age and type (general or psychiatric) were statistically significantly correlated with absenteeism for the first sample, and type with absenteeism for the second sample Absenteeism was negatively related to job satisfaction and positively related to psychological distress for the first sample only Well-being appeared to be a significant predictor of job satisfaction Song, Daly et al. 1997, USA To examine the effectiveness of a nurse-managed special care unit (SCU) for chronically critically ill patients, developed to improve patient outcomes as well as nurse outcomes.

Nurses n=143 of 208
Comparative study between two nursing units under two different models of practices. An annual survey was conducted among nurses in the two units. Study lasted 4 years.
The SCU nurses reported greater job satisfaction with payment and supervisory dimensions than the ICU nurses. The percentage of absent hours for nurses in the SCU was significantly lower than that for nurses in the ICUs Steenstra 2005, Netherlands To describe the course of sick leave due to LBP in a cohort of workers absent from work for 1 Health care professionals who reported sick leave due to nonspecific LBP 26 week prospective cohort study. After calling in sick, workers were obliged to fill in a Median time RTW was 5 days, median time LRTW was 6 days, median number of TDSL during follow-up was 6 days. day or more due to non-specific LBP.
To determine prognostic factors for duration of sick leave due to LBP.
for more than 1 day over a 2-year period n=615 sick leave questionnaire on the nature and expected prognosis of the sick leave episode. Case control study with data based on data from the Belgian cohort study of the relation between job stress and health called 'BELSTRESS' Sickness absence is measured in the year following the interview HCW score higher for perceived psychological workload but they experience less decision latitude in comparison with controls. For social support no significant differences were found. Among HCW there was a positive correlation between incidence of sickness absence and 'job demand'. The opposite was found among the controls For 'decision latitude' HCW reported in the highest quartile reduction in the odds of sickness absence. Controls showed a negative correlation between 'decision latitude' and sickness absence. Strong negative correlation between the incidence of sickness absence and 'social support' for HCW. Belonging to the 'job-strain' population increased risk of being absent Verhaeghe, Vlerick et al. 2006, Belgium To study how the occurrence and appraisal of recurrent changes (supervisor, tasks, colleagues, working hours, location) in the work environment of hospital nurses affect psychological well-being (i.e. job satisfaction, eustress and distress) and absence through illness.

Nurses n=2094
Cross-sectional research design. Absenteeism was measured in the 12 months following the questionnaire (measuring stress and satisfaction) Lower rates of sickness absence and higher levels of job satisfaction and job stress among male nurses. Nurses confronted with changes in the past 6 months noted higher levels of distress, no relation to job satisfaction and eustress, or to absenteeism. Strong association btn 'challenge' and job satisfaction, and eustress -no relationship with distress -no association sickness absence. The lower the level of 'threat', the higher the level of job satisfaction and eustress.
Positive association between the incidence of sickness absence and the level of 'threat' did not differentiate btn public and private Wier 1997, Canada To report on a randomized controlled trial that was designed to assess the potential for leadership effectiveness in improving staff morale and quality of care as well as its effect on the cost of nurse absenteeism.

Nurses and nurse managers
Consultative meetings with nurse leaders with the intention of influencing their leadership style and in return influence the absence rates of their teams. Improvement in the characteristics of their work setting and in the quality of working relationships, no statistically significant difference in absenteeism, Williams, Rondeau et al. 2007, Canada To see the impact that heavy workloads directly have on physician stress and more distally on physician job Physicians n=480 Initial survey was mailed to all physicians followed six weeks later with a second survey mailing.
Workload related to perceived level of Stress. Stress was negatively related to job satisfaction Job stress was found to be significantly satisfaction, job performance, patient care quality, absenteeism turnover intentions and organizational performance. To determine the degree of HH in the workplace and the extent that the perception of HH affect ill calls and the likelihood of leaving their current position

Nurses n=130
Retrospective descriptive crosssectional design 83.1% had seen HH. 95% believed that HH contributed to ill calls but only 20% admitted to having called in sick because of HH. 19% were definitely going to leave their current position and 20.5% were considering leaving their current position because of HH.