Article Text
PDF
Abstract
Objectives To identify diagnostic, patient/employee, general practitioner (GP) and practice factors associated with length of certified sickness episodes.
Methods Twelve-month collection of fit note data at 68 general practices in eight regions of England, Wales and Scotland between 2011 and 2013. Secondary analysis of sick note data collected at seven general practices in 2001/2002. All employed patients receiving at least one fit note at practices within the collection period were included in the study. Main study outcomes were certified sickness episodes lasting longer than 3, 6 and 12 weeks.
Results The data from seven practices contributing in 2013, and a decade previously, suggest that periods of long-term sickness absence may be falling overall (risk >12 weeks absence, OR=0.65) but the proportion of mild–moderate mental disorder-related (M-MMD) episodes is rising (26% to 38%). Over 32% (8064/25 078) of fit notes issued to working patients in the 68 practices were for a M-MMD. A total of 13 994 patient sickness ‘episodes’ were identified. Diagnostic category of episode, male patients, older patient age and higher social deprivation were significantly associated with the >3 week, >6 week and long-term (>12 week) outcomes, and GP partner status with the long-term outcome only.
Conclusions In the context of a rapidly changing legislative environment, the study used the largest sickness certification database constructed in the UK to enhance the evidence base relating to factors contributing to long-term work incapacity.
Statistics from Altmetric.com
What this paper adds
Long-term work incapacity has a detrimental effect on the well-being of the individual employee/patient. The longer the episode of sickness absence the less likely there will be a return to work.
Previous studies have failed to include diagnostic, patient, practitioner and organisational factors in the same explanatory model of long-term certified sickness absence.
This study, using sickness certification data collected at 68 general practices in the UK, presents some evidence that the ‘fit note’ introduced in 2010 is associated with a reduction in long-term certified sickness episodes.
Patient and diagnostic factors were found to be associated with the risk of having a long-term sickness episode. General practitioner and practice characteristics only marginally increased risk.
The findings are relevant to ongoing welfare policy developments in the UK, particularly the planned introduction of the Fit for Work Service to independently assess patients unable to work for health reasons.
Introduction
An estimated 131 million working days were lost in the UK in 2011 through sickness absence; over 2% of total working time and 4.5 days per worker.1 The associated direct costs in state benefit are £13 billion, plus £9 billion to employers in sickness benefits.2 Similar rates of absence are found in other European countries.2 Most episodes in the UK are relatively short (<4 weeks), with over 30% being self-certified by the employee (for the first 7 days of sickness). However, many episodes last longer than they need to, and an estimated 300 000 people move onto health-related benefits every year.2 The longer the person is out of work the more distanced from the labour market they become, reducing the potential for a return to paid employment.3
Concerns about the extent of long-term sickness absence (and the subsequent flow onto benefits) have resulted in two recent UK reviews into the health of the working-age population and the system of certifying sickness absence.2 ,4 Key recommendations introduced after the first review were to pilot case management services to reduce periods of sickness absence (Fit for Work Service (FFWS) Pilots), greater occupational health support for employers and general practitioners (GPs), and the introduction (in 2010) of a new universal Medical Statement of Fitness to Work, the ‘fit note’ (to replace the ‘sick note’). In the UK, up to 1 week's sickness absence is certified by the employee or ‘jobseeker’ for the employer or jobcentre, respectively; longer periods require a supporting medical statement from their GP. The fit note differs from the previous sick note in enabling GPs to include advice (to both the patient and employer) on how the patient may be able to return and perform some work with appropriate support (that may include phasing a return, amending duties, altering hours and/or making workplace adaptations). The shift to the new medical statement entailed a challenge to the long-held view that a patient/employee had to be 100% fit to return to work, and necessitated a recognition of the positive contribution work can make to an individual's health.5
The second review4 recommended the establishment of an Independent Assessment Service (IAS) to provide independent occupational assessment after 4 weeks of sickness absence to advise GPs and employers about work capacity and any associated ‘return to work’ support. The resulting Fit for Work Service (FfWS) is due to start operating in late 2014. It is anticipated that in addition to assessing and identifying the support required by the patient, the proposed service will help meet the needs of those GPs who feel that they lack the occupational health expertise or the time required to make informed decisions about work capacity. It will also benefit small/medium size employers unable to afford to purchase occupational health services for their employees.
This paper investigates the factors associated with the transition to medium-term and long-term certified sickness in the context of a new system of sickness certification and the proposed IAS. The availability of a large sickness certification data set provided a unique opportunity to investigate factors leading to work incapacity in a changing legislative environment.
The specific aims of the study were:
To investigate whether use of the fit note (compared to its predecessor, the ‘sick note’) has contributed to a reduction in the likelihood of the patient having a ‘long-term’ (>12 week) certified sickness episode at seven practices.
Using data from all 68 practices, to estimate the relative effects of a range of patient, GP and practice factors on the likelihood of the patient having a certified sickness episode exceeding (1) 3 weeks in duration (2) 6 weeks in duration and (3) 12 weeks in duration.
Methods
Data collection and processing
The study combines data from two projects commissioned and funded by the UK Department for Work and Pensions (DWP): the national evaluation of the fit note and the evaluation of FFWS pilots. The former recruited 49 general practices from five geographical areas of the UK (Scotland, Wales, Derbyshire, North West and South East England). The latter involved 19 practices sited in three FFWS pilot sites (Greater Manchester, Leicestershire and North Staffordshire). These practices varied in list size (small, medium and large), location (urban, suburban, rural) and deprivation. Ethical approval for data collection was obtained from the (UK) National Research Ethics Service in June 2011.
Although the evaluations had different objectives, both used a similar method of data collection, requesting GPs to use ‘carbonised’ pads of fit notes for a period of 12 months. Using the specialised pads ensured that details of every fit note issued in the period were retained on duplicate sheets. The FFWS evaluation practices started recording fit note data in July/August 2011 and the national fit note evaluation practices in November/December 2011.
A practice staff member anonymised and recorded details of each note on a dedicated spreadsheet. In addition to the details on the note (date of issue, diagnosis, period to abstain from work, whether the patient ‘may be fit’ to do some work, whether the patient needed to be re-assessed at the expiry of the note and the certifying GP) a number of additional items were collected from the patient practice record that have been shown to influence sickness absence risk. These included gender, year of birth and post code. The latter was transformed by practice staff into a neighbourhood deprivation score for the patient. Deprivation scores were based on lower-level Super Output Area and Data Zone scores in the most recent Indices of Multiple Deprivation for England, Wales and Scotland. Fit note data were transferred monthly by the practices to a project coordinator at the University of Liverpool who merged and cleaned the data. Post hoc transformations were conducted in order to allot individual diagnoses to one of 22 diagnostic groups, based on International Classification of Diseases Tenth Edition (ICD-10) categories (but modified to include common problems recorded on fit notes). On the basis of neighbourhood deprivation scores patients were allocated to five groups based on quintiles of lower Super Output Areas or Data Zones in their respective country. It was thus possible to establish whether the patient receiving the fit note was residing in one of the most or least deprived 20% of neighbourhoods in England, Wales or Scotland. At the end of data collection, practice managers were requested to provide some basic information about their certifying GPs (gender, age, partner status, full-time) and the practice itself (list size, location). Practices were assigned a deprivation status based on the proportion of their patients living in one of the most socially deprived areas in the country. Nineteen practices were classed as having a ‘very deprived’ patient population (ie, over 80% of patients living in the 40% most deprived neighbourhoods in the country).
Data from both studies, the national fit note and FFWS evaluations, were aggregated and analysed.
Main outcomes
A discrete certified sickness episode (of continuous certified sickness absence) lasting longer than 3 weeks (4 weeks of sickness absence, including the seven calendar days of patient self-certification).
A certified sickness episode lasting longer than 6 weeks.
A ‘long-term’ certified sickness episode exceeding 12 weeks in duration.
Data analysis
Practice ‘before and after’ analysis
For seven of the 49 practices in the national fit note evaluation corresponding comparative data from 2001 to 2002 were available. These data had been collected using the same ‘carbonised pad’ method to collate ‘sick note’ data at the practices over a full year. Hence, it was possible to compare the length of episodes certified by the fit note (2–3 years after the introduction of the new medical statement in 2010) and the sick note (8–9 years before the introduction of the fit note) in the same group of practices at two time points.
Multilevel, random-intercept logistic regression models were developed in order to estimate the independent effect of the fit note (compared to the sick note) in any reduction of patient risk of having a long-term certified sickness episode (after controlling for patient age, gender, social deprivation of lower Super Output Area and index diagnostic category of the episode). Separate models were run for each of the seven general practices, in addition to an estimate of overall effect. The OR, 95% CIs and p value are reported.
Cross-sectional analysis
Using data from all 68 practices a cross-sectional analysis was conducted in order to estimate the independent effects of a range of potential explanatory factors on the likelihood of patient certified sickness episode outcomes (>3, >6 and >12 weeks). These factors included patient/employee characteristics (gender, age and social deprivation), GP factors (gender, age, partner or not, part-time or full-time) and practice attributes (size, location and social deprivation status). The diagnostic category of the episode itself was also considered, along with whether the GP had included advice to the patient and employer about a potential return to work with appropriate support (‘may be fit’ advice).
Logistic regression models were developed in order to estimate the independent effect of each explanatory variable on likelihood of the patient having an certified sickness episode exceeding 3, 6 or over 12 weeks. In order to account for the hierarchical nature of the data (Episode/Patient/GP/Practice), appropriate multilevel random-intercept models were run. For each covariate in the fixed effects component of the model, the OR, 95% CIs and p value are reported.
For all models in both analyses, a conventional criterion of statistical significance (p<0.05) is used.
Only fit notes issued to patients known to be absent from usual paid employment (ie. recorded as ‘in work’) were included in the cross-sectional analysis. All data were analysed using SPSS for Windows V.20 and STATA V.10.
Results
‘Sick note’ and ‘fit note’ compared
The availability of sickness certification data collected from seven practices at two different time points allowed an assessment of how effective the fit note was in reducing long-term patient certified sickness episodes, compared to its predecessor, the sick note.
Table 1 reports the proportions of medical statements and certified episodes allocated to diagnostic categories in the 2002 and 2013 studies at the seven practices. (Fit note and episode figures for all 68 practices in 2013 are included in this table as an indicator of the degree of ‘representativeness’ of the seven practices). By the time of the second study, after the replacement of the sick note by the fit note, mild–moderate mental disorders (M-MMD, such as anxiety, depression and stress) had increased substantially as a cause of certified sickness absence (table 1). While in 2002, this type of common mental health problem was cited on 32% of sick notes and was mainly responsible for 26% of certified sickness episodes at the seven practices, the corresponding figures for 2013 had risen to 41% and 38%. The other notable change was the reduction in the prevalence of respiratory health problems requiring certification (down from 10% to 6% of total fit notes).
Diagnostic categories of medical statements and certified sickness episodes at seven practices in 2002 and 2013 (including all patients, regardless of whether employment status was recorded or not)
Long-term (>12 week) episodes certified by the sick note and fit note
When patients from the seven practices were pooled in multilevel logistic regression analysis there was a highly significant overall reduction in the risk of a long-term (>12 weeks) certified sickness episode in the later fit note study (OR=0.65, 95% CIs 0.58 to 0.72, p<0.001). However, the effect varied between individual practices.
At three of the practices, after adjustment for patient and diagnostic factors, the likelihood of a patient having a long-term episode was significantly reduced in the later study using the fit note, compared to the earlier sick note study. In the first of these practices, there was a 61% reduction in risk of a long-term certified sickness episode in the 2013 study (adjusted OR=0.39, 95% CIs 0.31 to 0.51, p<0.001), in the second there was a 49% reduction in risk (adjusted OR=0.51, 95% CIs 0.40 to 0.66, p<0.001) and a 36% reduction in the third practice (OR=0.64, 95% CIs 0.45 0.92, p=0.02). In the remaining four practices, there was no evidence of a statistically significant increase or decrease in long-term episodes in the later fit note study.
Fit notes issued to working patients at the 68 practices
The 68 practices submitted details of 79 815 fit notes issued to 33 768 patients in the data collection period. However, a record of the fit note recipient being ‘in paid employment’ was only available for 10 984 patients (32.5% of total). For 20 544 (60.8%) patients receiving a fit note it was not possible to establish whether they were usually in paid employment or not. Only the fit note recipients known to have been usually in work were included in these analyses. This employed subgroup received a total of 25 061 fit notes, accounting for 13 694 certified sickness episodes (of continuous certified sickness absence). The rate of issue of fit notes was 2.3 per person year. This group of patients recorded as ‘in work’ did not significantly differ from the excluded patients (whose employment status was not known), in terms of gender, age, social deprivation and diagnostic category of sickness certification.
In relation to both fit notes and patient episodes the diagnostic category accounting for the highest proportion of certified sickness was the M-MMD category (that included common psychological causes of sickness absence such as ‘depression’, anxiety’ and ‘stress’; table 2). This category accounted for 32% of all fit notes issued to employed patients in the database and 29% of all certified sickness episodes. Back problems were also an important cause of certified sickness, accounting for over 10% of fit notes and nearly 9% of episodes. It is not uncommon for the GP to cite a symptom (eg, ‘pain in…’) rather than a specific diagnosis as the reason why the patient should abstain from work. This ‘symptom’ category accounted for nearly 10% of all fit notes and episodes in the database. Nine per cent of fit notes were issued, and a similar proportion of sickness episodes certified, in order to enable a patient to recover from a recent surgical operation.
Fit notes* and certified sickness episodes† within diagnostic categories (including only patients at the 68 practices recorded as ‘in work’)
The fit note method of certification enables the GP to advise the patient (and employer) that, with appropriate support, the patient ‘may be fit’ to perform some of their work responsibilities. A total of 2151 fit notes issued to recorded employed persons, representing 8.5% of all notes to this subgroup, had the ‘may be fit’ box checked. Of the four structured advice items on the fit note, ‘amended duties’ was most often indicated. ‘Amended duties’ was ticked on 955 fit notes (44% of all ‘may be fit’ notes to employed patients), either alone or along with other items of advice. Proportions of ‘may be fit’ notes indicating ‘phased return’, ‘altered hours’ and ‘workplace adaptations’ were 34%, 20% and 9% respectively.
A total of 1602 (11.7%) certified sickness episodes were concluded by a fit note advising that the patient may be fit to return to work with appropriate support (ie, having one or more of the advice items indicated by the GP). Such ‘may be fit’ advice was more prevalent within the non-mental health categories of episode. Over 17% (210/1205) of certified sickness episodes due to a fracture or other injury and 15.9% (295/1858) of back or other musculoskeletal episodes included this return to work advice. Sixteen per cent (209/1314) of episodes certified because of the need for rehabilitation after a surgical operation included advice that the patient ‘may be fit’ to do some work. Only 9.9% (393/3950) of certified sickness episodes within the M-MMD category concluded with return to work advice from the GP.
Length of certified sickness episodes
Nearly 53% of all certified sickness episodes did not exceed 3 weeks in duration. Over 35% were longer than 3 weeks but did not exceed 12 weeks in duration and nearly 12% were classed as long-term (lasting longer than 12 weeks; table 3). The duration of episodes varied widely between diagnostic categories. Certified sickness episodes in the respiratory category, which largely included fit note certification due to acute upper respiratory infections, tended to be short (85% of all episodes within this category were for 3 weeks or less). Digestive, infection, genitourinary, skin and pregnancy-related sickness certification tended to be relatively short (over 65% of certified sickness episodes in these categories not exceeding the 3 week threshold). Where only a ‘cause of injury’ (eg, ‘Road Traffic Accident’, ‘Fall’, ‘Assault’) rather than the injury itself was specified, the resultant certified sickness episode was more likely to be short.
Duration of certified sickness episode by index diagnostic category (including only patients at the 68 practices recorded as ‘in work’)
On the other hand, some health problems were more likely than others to result in long-term certified sickness episodes. The severe mental disorder and congenital categories contained relatively high proportions of episodes exceeding 12 weeks in duration (24% and 56% of episodes, respectively, within the two categories). Forty-nine per cent of neoplasm/cancer episodes were for longer than 12 weeks. Over a quarter (27%) of circulatory illness-related episodes were long-term in duration. Certified sickness episodes for a musculoskeletal problem (affecting bodily sites other than the back) had a sizeable proportion (22%) of episodes exceeding 12 weeks. A majority of these longer musculoskeletal episodes involved work incapacity due to chronic arthritic conditions.
Patient, GP and practice factors associated with episode outcomes
After running three separate multilevel logistic regression models, a number of factors were found to be associated with significant increases in the risk of sickness absence persisting for 3, 6 and 12 weeks or more (table 4). To measure effects of diagnostic category on outcomes, the respiratory problem category (with the highest number of short episodes) was selected as the reference. The significant diagnostic categories (in descending order of impact) were mental health problems (3 week, OR=7.03; 6 weeks, OR=6.39; 12 weeks, OR=4.33), circulatory problems (3 weeks, OR=6.51; 6 weeks, OR=9.53; 12 weeks, OR=9.30) and musculoskeletal problems (3 weeks, OR=4.39; 6 weeks, OR=4.93; 12 weeks, OR=4.67). Other significant factors were: age (increased risk per 10 years age band, OR=1.19, 1.09 and 1.03 at 3, 6 and 12 weeks, respectively), living in the most deprived quintile of areas (OR=1.24, 1.23, 1.24 at 3, 6 and 12 weeks) and being male (OR=1.18, 1.30. 1.35 at 3, 6 and 12 weeks). A final fit note in the certified sickness episode signifying the claimant may be fit for some work was associated with a higher risk of being incapacitated for over 3 weeks (OR=1.89) and 6 weeks (OR=1.61) but a reduced likelihood of a long-term episode (OR=0.72) The only GP or practice factors significantly associated with any outcome were being certified by a GP partner (compared to a salaried or locum GP), which increased risk for a long-term episode by 25% (independent of other factors such as age of the GP); and practices delivering care to the highest areas of deprivation (OR=1.18 at 3 weeks, compared to practices with less deprived patient populations).
Patient/employee-level, GP-level and general practice-level factors associated with certified sickness episodes exceeding (1) 3 weeks (2) 6 weeks and (3) 12 weeks in duration (including only patients at the 68 practices recorded as ‘in work’)
Discussion
Main findings
Our study confirms that, in the UK, common mental health problems (such as depression, anxiety and stress) continue to account for a high proportion of total certified sickness, both in the form of individual fit notes issued to the patient and certified sickness episodes. This contrasts with other countries, where musculoskeletal problems remain the largest category, and historical trends in sickness certification within the UK itself.6 ,7 Across 68 practices, approximately a third of certified sickness was attributed to these M-MMD. In the subanalysis, based on seven practices collecting sick note data in 2001–2002 and fit note data in 2011–2013, there was further evidence of the increasing importance of psychological health problems. Episodes due to this type of mental health problem increased from approximately a quarter to a third of all certified sickness episodes at the practices over the intervening decade.
In terms of duration, over 47% of all episodes were for a period exceeding 3 weeks, and more importantly (given the potential consequences of long-term work incapacity), 12% of episodes were long-term (>12 weeks in duration). In terms of the fit note, there is a suggestion that it may have helped significantly reduce absence episodes in at least some of our seven comparison practices. The larger study confirms our earlier identification of key risk factors for longer absences, namely diagnosis, deprivation, increasing age and being male. Having a ‘may be fit’ fit note issued by the GP (advising that the patient may be able to do some work) was significantly (and negatively) associated with likelihood of a long-term certified sickness episode. The only GP or practice factor associated with the long-term outcome was the certifying GP being a partner.
Strengths and weaknesses of the study
The major strength of the study is that the database on which it is based, including data collected from two separate evaluations, represents the most comprehensive collection of sickness certification ever conducted in the UK. The data are representative, having been collected from general practices in eight distinct regions or counties of the UK, for a period of 12 months (accounting for any seasonal variation in issue of fit notes). The participating practices themselves varied in terms of size (patient list and number of GPs), location and social deprivation status. In sum, the database provides a valuable source of evidence to support research and inform policy, which, apart from our earlier study, has continued to rely on reported behaviour rather than primary evidence. The study was limited by the range of routine data that was able to be collected. It would have been useful to have been able to measure more patient-based variables relating to employment/occupational status, individual socioeconomic status (such as housing tenure), and domestic circumstances (such as marital status and number of children). The lack of such basic information as to whether the patient in receipt of the fit note was actually in work at the time reduced the numbers that could be included in our analysis. However, such data are not collected systematically in patient records at general practices in the UK, nor on the fit note itself.
In the analysis comparing the fit note with the previous sick note the degree of specification error is also a consideration. The exclusion of potentially key explanatory covariates (such as patient occupational status, home circumstances, job control and satisfaction) may have exaggerated the impact of fit note use. Other weaknesses in design and analysis include the lengthy interim period (over 10 years) between sick note and fit note data collection, the high turnover in GPs at the practices (only 60% of GPs issued medical statements in both collection periods) and unrecorded changes in composition of patient populations during the interim period. For patients whose sick note data was collected in 2001–2002 no additional information relating to employment status was recorded.
Relation to other studies
A number of previous international studies have found that patient gender, age, deprivation and diagnosis are strongly associated with the risk of sickness absence (or ‘sicklisting’) and the length of the resulting episode.8–21 Our study provides robust evidence that these remain important risk factors in the UK. Other factors found elsewhere include family circumstances22 ,23 and type of work,24–26 but the UK sickness recording system does not allow these factors to be systematically studied. Relatively few studies have looked at the characteristics of the GP that may influence the length of sickness absence (after controlling for patient factors). There is limited evidence suggesting GP age, gender, salary and workload may influence outcome.9 ,27 ,28 However, confirming results of our previous research,29 we found that the combined effect of the measured GP factors on certified sickness outcomes was relatively unimportant (compared to diagnostic and patient factors), which concurs with the findings of a recent study utilising a multilevel analytical model and a continuous outcome.9 We did find that the GP being a partner (rather than a salaried GP or locum) significantly increased the likelihood of a patient having a long-term certified sickness episode, regardless of patient or diagnostic factors. Methodological factors may partly explain this association. In episodes where more than one GP had issued constituent fit notes, the ‘certifying GP’ status was assigned to the GP who issued the majority of notes. It is also likely that a partner in a practice would be more likely to see the patient throughout the duration of the episode. A combination of these two factors, methodological and ‘continuity of care’, may explain why this GP variable was associated with the long-term (>12 weeks) outcome. Organisational factors about general practices have not been considered in previous research. However, we found that one practice characteristic, the social deprivation status of the practice patient population, was a significant independent predictor of the >3 week outcome (but did not retain significance at 6 and 12 weeks, which may be because within-practice population variables become more important at that stage).
Implications for health and welfare policy in UK
One of the anticipated benefits of the fit note was to return more patients to work earlier in the certified sickness episode, and to contribute to an overall reduction in long-term work incapacity. This was one of the criteria tested in the national evaluation of the fit note.30 Our analysis of the fit note data provides some tentative evidence that this objective may be being met. Although less than 12% of all episodes included some GP advice (to the patient and employer) that the patient may be able to return to work with adequate support, this option was significantly less likely to be used when the episode had become ‘long-term’ (>12 weeks), which may be because those absent for over 3 months are the least able to work in the foreseeable future and partial work is not recognised as an option. Furthermore, there is an overall downward trend in long-term episodes in the seven practices included in the 2002 and 2013 comparison, significantly so in three of them.
It is proposed that a new service, the FfWS, provided by the UK's largest independent occupational health provider, will be introduced in late 2014 in order to be able to independently evaluate patients who have completed a fourth week of sickness absence. We found that 47% of episodes in our study lasted longer than 3 weeks (4 weeks, including 7 days of self-certification). Even if only a relatively small proportion of this group of working patients had been offered, and had consented to an independent assessment, it would still require considerable resources (in terms of recruiting occupational health doctors, and training GPs and nurses) in order to effectively operate such a service. Our findings suggest that a more useful strategy might be for GPs to target certain ‘at risk’ groups for more in-depth discussion about work and health.
The large-scale Improving Access to Psychological Therapies (IAPT) programme aims to significantly increase the availability of evidence-based psychological treatments for common mental health problems (particularly anxiety and depression) within the National Health Service in England. The first wave of IAPT services was implemented in 2008. At the outset it was anticipated that the IAPT programme would eventually ‘pay for itself’ by generating net savings for the Exchequer, in reducing state spending on sickness benefits, and for employers, in reducing the number of sickness absences caused by psychological morbidity.31 ,32 The 2012 review of the first 3 years of IAPT claimed that 45 000 people using a IAPT service had been assisted to move off sick pay and benefits.33 However, the results of our study indicate that M-MMD still account for about a third of all certified sickness absence, and this proportion may well have increased since the introduction of IAPT. It is of course possible that the burden of sickness absence due to these problems may have been even greater without the availability of IAPT. However, our findings do suggest that the optimism expressed in the review of IAPT achievements may have been somewhat overstated.
Study findings in a wider international context
It is very difficult to place our study findings, relating to types and length of certified sickness, in a wider context by comparing them with those of similar studies conducted in other countries. Attempting to compare sickness absence rates between countries is a complicated process because of the use of different definitions of ‘absence’, variations in criteria for inclusion and the different populations and economies being compared. A ‘league table’ of countries’ sickness absence rate would be potentially misleading.34 Also, while we have classed a sickness absence episode as ‘long-term’ on the basis of a threshold (>12 weeks) inferred from empirical data, other countries in Europe, with different administrative requirements, have used ‘long-term’ to categorise a period of sickness absence exceeding thresholds ranging from 6 weeks to 36 weeks.35
One of the few attempts to compare sickness certification rates across national boundaries was made in a 2008 systematic review that retained 11 studies in six European countries.36 The review standardised measures of certification on the basis of the rate of issuing medical certificates (number of certificates per person). While a large variation in rates was reported, the authors concluded that the variation may be largely explained by differences in the period of ‘self-certification’ permitted and the system by which sickness benefits were paid. The UK system stipulates that the sick employee should wait 7 days before requesting a fit note from a GP, while in Malta there is a requirement to obtain a certificate on the first day of sickness. In the UK, the employer is usually responsible for paying the sick employee for the first 28 weeks (if sickness certificates are provided), while Norwegian employers only pay the first 16 days before the National Insurance Administration continues payment. It would be anticipated that the shorter the required period of self-certification and the longer that employers have to pay absent employees, the higher the rate of issuing certificates.
Outside the UK, systems of sickness certification typically still rely on an absolute judgment of whether a person is ‘fit for work’ or ‘not fit’. As already reported the introduction of the fit note in the UK in 2010 followed an increasing awareness that performing some work was beneficial to health and it was important to prevent transition to long-term work incapacity. In fact, a similar rationale has been used in the implementation of the system of ‘partial sick-listing’ that has been available to physicians in the Nordic countries (Sweden, Norway, Denmark and Finland) for a number of years. The patient can return to work on a part-time basis, or work full-time with reduced work duties, and still receive partial sickness benefit in addition to a part of their salary. The few studies that have examined partial sick-listing have reported findings that correspond to ours, specifically that non-mental health problems were the most common problems to result in partial sicklisting.37 In all four Nordic countries that examined partial sick-listing, poor collaboration among the interested parties (ie, employer, employee, physician and insurance institution) prevented optimum use of the partial sick leave programme.
Our study reported that ‘may be fit’ advice was more prevalent on fit notes issued for physical health problems. This implies that this type of sickness certificate may be particularly effective in economies where a large proportion of the labour force still work in manual occupations. A 2007 survey of working conditions in 31 European countries requested respondents to provide information on recent sickness absence.38 In some countries in central and eastern Europe (Poland, Slovakia and Slovenia) physical health problems were responsible for a high proportion of total reported sickness absence, while mental health-related absence was relatively insignificant. However, in countries with a large service sector (UK, Germany, Netherlands, Austria) physical health-related sickness absence only accounted for a minority of all absence. Outside the European policy arena, the Royal Australian College of General Practitioners have welcomed plans for a trial of a new sickness certificate that may lead to adoption of a fit note similar to that introduced in the UK in 2010.
Implications for future research
Although this study (and our previous research) has confirmed the importance of diagnostic and patient factors in any explanation of long-term sickness certification, there is an urgent need to extend the scope of research via the consideration of other potential explanatory factors. Such consideration would necessitate UK researchers obtaining access to data other than that routinely recorded in patient records. For instance, information about the patient's employment status (employed, or unemployed and claiming benefits) and usual occupation would be a minimum requirement, along with a measure of individual GP attitudes to certifying sickness. An attempt should be made to collect such data, either in primary research or using any general practices that may routinely record such additional data.
Acknowledgments
The authors would like to acknowledge the efficiency of clinical and administrative staff at the 68 practices collecting data for the study. Also, the authors are grateful to Jane Rees and Tom Walley for their helpful comments on provisional drafts of the paper.
References
Footnotes
Contributors CS assisted in data collection, was responsible for data analysis, and compiled a first draft of the paper. MG led the study, provided clinical expertise and revised subsequent drafts of the paper. JH coordinated data collection and contributed to revision of content and formatting.
Funding The two national evaluations providing source data for the study were both funded by the UK Department for Work and Pensions. The funding agency did not have any role in the study design, analysis or interpretation of data, or in the decision to submit this article for publication. Hence, authors were acting fully independent of funders.
Competing interests None.
Ethics approval The construction of a national fit note database was approved by the National Research Ethics Service (NRES) and National Information Governance Board (NIGB).
Provenance and peer review Not commissioned; externally peer reviewed.