In this study, we examined, first, the characteristics of frequent attenders (FAs) in three Finnish outpatient health care schemes and, second, the association of frequent attendance with sickness absence in the working-age population of one Finnish city. We assessed frequent attendance in the public, OHS and private care schemes. The FAs, here the top decile of attenders, accounted for over one third of all included attendance days. The disproportional use was clearest in public care, but the large cumulative consumption of the service capacity by a small number of clients was found in every scheme.
Characteristics associated with frequent attendance
As in previous studies [6, 8, 9, 11, 12, 15,16,17,18,19], frequent attendance was most characteristic for females, older age groups, and those with a lower socioeconomic status. However, especially the socioeconomic characteristics of FAs depended on health care scheme, reflecting the different roles of the schemes in the Finnish health care system. In the public scheme especially, frequent attendance was common among those outside work or with only basic education. In general, lower socioeconomic groups in Finland are known to use mainly public services [26, 47]. Frequent attendance in the public scheme was also common for individuals with a chronic disease, and especially among those with multiple diseases. This may partly result from the tendency to refer long-term conditions to public care. The capacity of the public care scheme is thus heavily consumed especially by more financially vulnerable individuals and those with multi-morbidity. Although not studied here, social problems and negative life events are prevalent in these groups [5,6,7, 10]. These individuals are also often outside the scope of OHS services in Finland and in many cases cannot afford private care.
In the OHS scheme, frequent attendance was most common among lower non-manual and manual employees. Among employees, those with a lower socioeconomic position are most likely to use OHS, especially when adjusting for age and chronic disease [17, 26]. This generally higher attendance rate may be reflected in the higher proportion of FAs as well. Musculoskeletal disorders are both clearly more frequent [48] and pose a greater risk for disability retirement [49] among lower non-manual and manual employees than among upper non-manual employees. These disorders probably account for the higher FA rate in the OHS scheme as well. In addition, for some lower non-manual and manual employees frequent attendance may be caused by the need for a physician’s certificate from the first absence day. This kind of policy is more common for lower occupational classes in Finland [50]. Older age was associated with frequent attendance especially in OHS. In addition to the increasing incidence of health problems with higher age, this probably reflects the comprehensive services commonly available through OHS. OHS are available to most working-age Finns free of charge at the point of delivery, and employees with health problems have quite generous access to them according to medical need.
In private care, frequent attendance was particularly common among entrepreneurs. Obtaining OHS coverage is voluntary for entrepreneurs, and only 43% were estimated to do so in 2016 [51]. Thus, many entrepreneurs in great need of care may complement public care with private care.
Attendance frequency and sickness absence days
Our main contribution is to shed light on the association between frequent health care attendance and subsequent sickness absence (SA) in the three health care schemes. FAs of each scheme had on average significantly more SA days than others in 2016 and this difference, albeit somewhat diminished, remained for the two follow-up years. The fact that the association between frequent health care attendance and SA days was strongest in 2016 can be explained by several reasons: One, because health care attendance was also measured for that year; two, because a physician’s sickness certificate is a prerequisite for sickness allowance; and three, because the dynamics between attendance and SA can work both ways. In the OHS scheme, the diminishing association over follow-up years may also be associated with adjusting the workload to better fit the individual capability of some employees. On the other hand, the association diminished similarly in all schemes. In all, frequent outpatient health care attendance, regardless of scheme, was a predictor of sickness absence for several prospective years.
Frequent attendance was associated with SA in all schemes but the strongest association was seen with public care. Combined with the lower socioeconomic status of public care FAs, the results indicate that especially the public care FAs have recurrent consultations without finding a solution to their primary problems, leading to a sick-leave later. This vulnerable group can include those with cumulative medically unexplained symptoms, health anxiety, negative life events and social problems [5,6,7,8,9,10]. Adjusted for covariates, public care FAs had on average 62 sickness allowance days in 2016–2018, reflecting the unmet needs and severity of health issues in this group. Interestingly, adjusting for covariates lowered the effect only little. This indicates that while an individual’s background affects the probability for frequent health care attendance, it does not appreciably shield individuals against the negative outcomes of excess attendance.
Further, our study shows that the association between frequent attendance and SA has different diagnostic emphases in different schemes. This demonstrates that studying the association in one scheme can lead to biased conclusions. Even though the risk for SA due to all three diagnostic groups was prominent for FAs in the public scheme, frequent attendance increased the expected number of SA days especially due to mental disorders. Individuals with low socioeconomic status are known to suffer from mental disorders more often than other socioeconomic groups [52]. As they are also more often public scheme FAs, it is understandable that their frequent attendance can later be accompanied by sickness absence due to mental disorders. The fact that adjusting for covariates reduced the effect that frequent attendance in the public scheme had for mental disorders especially suggests that this mechanism is true. In addition, a combination of comorbidity, social problems and negative life events may weaken public scheme FAs’ resources for coping with the initial mental health issues, possibly leading to future SA days. In our study the entitlement to reimbursed medicines based on multiple diseases - a marker for comorbidity- did indeed increase the probability for being a FA in the public scheme.
In OHS the association between frequent attendance and SA was strong especially for SA days due to musculoskeletal disorders. This is in line with Reho’s studies in which alongside mental disorders, especially musculoskeletal disorders as a consultation or a sickness allowance diagnosis have been associated with FAs in OHS [3, 4, 20]. Partly, this scheme-specific effect may reflect the use of the scheme in general. Musculoskeletal diseases and mental disorders are the main work-related reasons for OHS visits in Finland [53]. In private care, the association between frequent attendance and SA was not as strong as in the other two schemes, not even in the unadjusted models. This indicates that health issues are in general less severe for FAs of the private scheme than FAs of other schemes. Medical conditions might also be treated more quickly in self-paid private care, as the patient does not have to queue and doesn’t need a referral from primary care to specialized care provided in private care.
Strengths and limitations
Our data on outpatient health care are exceptionally extensive, including all schemes relevant to the Finnish working-age population. To our knowledge, no studies on frequent attendance and its associations with SA with equivalent registers have been published. Our study included all residents aged 25–64 of the city of Oulu, and was based fully on register data deemed to be highly reliable and objective, with very little missing information, no self-report bias and no loss to follow-up. Furthermore, we were able to reliably calculate both the number of attendance days in each health care scheme and the precise length of sickness absence spells.
However, there are also some limitations. The findings do not cover the whole working-age population and are restricted to outpatient care only. We could not specify the proportions of primary and secondary care or the proportions of visits to the various professionals. This warrants caution in comparing the schemes. Due to the observational nature of the study, no causal effects can be shown between health care attendance and SA days. Naturally, the association does not run one way, but can work in both directions. For instance, an initial period of SA may be followed by rehabilitation or specialist care. However, it is probable that using health care affects the receipt of SA benefits. Health issues are mainly dealt with in health care, and only secondarily do they lead to long-term sickness absence. Also, a physician’s certificate is required to begin a period of sickness allowance. In future studies, both health care attendance and sickness absence should be followed over several years. Reho et al. [20] and Smits et al. [5] have compared occasional (1 year) and persistent frequent attendance, but a true longitudinal research setting is still lacking.
Finally, it has to be noted that the definition of FA as the top decile of individuals who attended outpatient health care can have implications for the results concerning attendance frequency and sickness absence days. Although FAs were defined similarly across all schemes in relative terms, the number of visits required to be classified as FA varied by scheme. As a sensitivity analysis we alternatively defined FAs in each scheme as individuals who attended health care on 10 days or more. The analysis did not change our key results: The effects (IRRs) of frequent attendance on SA were still strongest in the public scheme. Moreover, frequent attendance increased the SA days in the public scheme especially due to mental disorders, and in OHS especially due to musculoskeletal disorders. However, the effects for FAs strengthened in the private scheme in general, and frequent attendance now increased the SA days especially due to mental disorders. It is noteworthy however that the alternative definition of FAs used in the sensitivity analysis is problematic. As a result of standardizing the cut-off point of (ten) absolute attendance days, the relative proportions of FAs obviously diverged, respectively: in the private scheme, FAs comprised only 0.4%, but in all outpatient health care 25% of the total study population. We conclude that the decile-based relative method for defining FAs in different schemes is well-founded, and adequate in its ability to discriminate between groups [30].
Practical implications
FAs are a vulnerable group [5,6,7,8,9,10] for whom poor health but also negative life events can be strong predictors of long-term sickness absence [7]. As the group may attempt to address unmet clinical and social needs with excess use of health care, they should be targeted with a wider service palette, including the possibility to access social services.
Many national systems have aimed for models of healthcare that can better coordinate and integrate various services, aiming to reduce fragmentation and add continuity of care (e.g. [54]). However, targeted interventions to influence morbidity, quality of life, and healthcare utilization among FAs have been modest in their effectiveness [55]. One method proposed is a personal case manager to ensure, coordinate, and integrate services for the patient [56]. A similar personal coordinator has been proved efficient in reducing public health centre visits among multimorbid clients [57]. A further key target for FAs could be to use frequent health care attendance as a marker for an assessment of need for rehabilitation. The fact that frequent attendance is a predictor for disability pension [58] supports such a targeted step.
Our study shows that the problem domain related to frequent attenders can be found in each health care scheme, and that there are FAs consuming capacity from multiple schemes as well. The strongest association with future SA days was found in this group. They cannot be identified without systematic patient registers shared by various schemes. The shared data base should include all health care, as well as social care visits regardless of scheme. In Finland, a recently developed electronic database, called Kanta services, makes sharing patient information (with a consent from a client) between health care providers possible. This may contribute to the continuity of care as well. In OHS, the continuity of care might be improved further by systematically guiding FAs to familiar OHS professionals.
More generally, the fact that lower socioeconomic groups are overrepresented among FAs showcases the strong link between socioeconomic status and health (e.g. [59, 60]). However it also indicates that polarization of health care attendance is not immune to change. Actively providing preventive care to the population with lower socioeconomic status may decrease cumulative somatic, mental and social problems, and therefore decrease the risk for excessive health care use later. As for employees, work environment factors such as occupational risks, physical and mental workload, and possibilities to influence one’s work affect the amount of OHS health care attendance [25] alongside the characteristics studied here. Our study showed that older employees, lower non-manual employees and manual workers are most often FAs in OHS, implying the potential of workplace adjustments for these employee groups.