Long term work disability, characterized by a prolonged inability to work due to health related-reasons, is a significant public health problem, affecting both the individual (Thomas, 2005; Waddell et al., 2006) and society at large (OECD, 2019). It not only affects the economic standing, social well-being, and overall health of individuals (Thomas, 2005; Waddell et al., 2006), but also serves as a strong predictor of permanent work life exclusion (Gjesdal & Bratberg, 2003; Hultin et al., 2012; Klein et al., 2021; Lund et al., 2008). Shortening the duration of benefits can be cost-effective from both an employer and societal standpoint, through enhancing productivity while reducing lost income and tax revenue.

The Scandinavian countries of Norway, Sweden and Denmark are known for their robust social security and sickness insurance systems, providing comprehensive support to individuals unable to work due to illness or injury (Pedersen & Kuhnle, 2017). Local social insurance offices, such as the Norwegian “NAV” (the Norwegian Labour and Welfare Administration), hold key positions in facilitating the return to work (RTW) process for individuals attempting to resume employment following illness or injury (Hakvaag et al., 2022). Through administrating sickness benefits and offering and monitoring activities aimed at improving RTW outcomes such as duration of benefit recipiency and number of hours worked, social insurance offices hold an important coordinating role in the RTW process.

In NAV, the public employment service, the national sickness insurance, and the social services are co-located, and the central aim of the organization is to provide holistic and individualized services (Christensen & Lægreid, 2011). In Norway, sick-listed workers can receive sickness benefits for up to a year (52 weeks), and if unable to RTW after this period, or if individuals i.e., due to income requirements, are not eligible for sickness benefit they can apply for Work Assessment Allowance (WAA called Arbeidsavklaringspenger or AAP in Norwegian (NAV, n.d.-b)). WAA is a temporary health-related benefit lasting up to three years for individuals with certified illnesses or injuries affecting their work ability. Following a government reform in 2018 the maximum duration for receiving WAA benefit was reduced from four to three years. This legislative change, combined with other measures to narrow the entrance of individuals to the benefit and increase work incentives, was intended to facilitate faster and more individually tailored support for the recipients (Prop. 74L, 2016).

The overall aim of the WAA benefit is to clarify service users’ employment prerequisites and the WAA spell can result in full or partial transition to work and/or permanent disability pension. Participation in the WAA program requires active medical treatment or engagement in activation programs aiming to clarify the individuals’ workability or to promote RTW (NAV, n.d.-b). These activities, organized by local NAV offices or external providers e.g., include job training, wage subsidies, shorter labor marked courses to improve individual qualifications and employment opportunities, and education (NAV, n.d.-a). Common for the activation programs is that the NAV advisor is responsible for assessing the service user’s follow-up needs and initiate necessary and beneficial measures to promote RTW.

The duration of the work disability and the benefit duration is an important parameter that can affect the likelihood of returning to work (Gjesdal & Bratberg, 2003; Hultin et al., 2012; Lund et al., 2008). Despite this, little is known about the impact of the follow-up from the local social insurance office (in Norway: the local NAV office) on service users’ RTW outcomes. Norway, Sweden, and Denmark are known for having advanced systems for collecting and managing administrative data, which provide detailed insights into service users’ benefit trajectories over time (DREAM, n.d.; Forløpsdatabasen-Trygd, 2002; Statistikmyndigheten SCB, n.d.). However, effectively utilizing this data to assess the impact of RTW interventions, as well as the follow-up from local social insurance offices, is challenging due to the inherent complexity of the RTW process. Challenges such as risk of reversed causation and selection issues make it difficult to accurately estimate the effectiveness of RTW interventions. This highlights the importance of understanding the specific policy context and how interventions are implemented.

Our study extends the existing literature by investigating inter-office variations in service users’ outcomes within local NAV offices in Rogaland County. The study applies a generalized linear regression model analysis to investigate the association between local NAV office affiliation and service users’ benefit duration within the same social policy context and labor market region. A retrospective observational design was employed to investigate the characteristics of seven local NAV offices within the Nord-Jæren region in Norway. The study utilized individual-level data (n= 8,582), aggregated to office level, to follow service users through former disability spells. This approach allowed for a comprehensive analysis of WAA benefit duration, and to identify characteristics of the local NAV offices’ practices related to WAA duration and the use of activation measures within the local NAV-offices.

Loisel et al (2005) provides a comprehensive framework for understanding the factors that influence RTW, emphasising that work ability and RTW are not solely dependent on the individual and his or her individual characteristics, but are also influenced by workplace conditions, healthcare providers, the insurance context, and the collaboration among these stakeholders. Through the administration of compensation for lost income, the insurance context can exert a significant influence on the RTW process (Loisel et al., 2005). Moreover, systemic factors and administrative practices can play a pivotal role in shaping the experiences and outcomes of individuals attempting to RTW (Collie et al., 2019; Frøyand et al., 2018; Gray et al., 2019; Hakvaag et al., 2022; Lane et al., 2022; Lippel, 2007; MacEachen et al., 2010).

Although centrally governed by the Directorate of Labour, local NAV offices differ across various parameters, including the office size (number of employees), caseload, competence, and organization of work and practices (Fossestøl et al., 2016; Langeland & Galaasen, 2014; Proba, 2015). These variations in the NAV offices organizational settings can shape service delivery (Jewell & Glaser, 2006), and influence outcomes such as benefit duration for service users. Investigating inter-office variations in service users’ outcomes have the potential to reveal unwarranted variations in benefit duration. By identifying variation that cannot be explained by individual-level factors and examining potential office-specific characteristics or organizational practices associated with longer WAA durations, policy makers and practitioners can pinpoint areas requiring interventions. Addressing unwarranted variation can enhance equity and fair treatment in public service delivery, both within and between regions. Additionally, addressing unwarranted variations can lead to improved outcomes for individuals seeking to RTW and potential cost savings.

In total, 8,582 service users received WAA benefits and follow-up services from the seven NAV offices during the five-year follow-up period. Of the service users, 41.0% were male, and 59.0% were female. The sample predominantly comprised Norwegian service users, with 79.8% having one or two Norwegian-born parents. Regarding age distribution, the majority fell into one of three age-groups: 30 – 39 years (20.6%), 40 – 49 years (24.0%), and 50 – 59 years (25.8%). The two age categories, 18 – 29 years (16.3%) and 60 – 67 years (13.3%), encompassed the lowest proportion of service users. The NAV offices varied in terms of office size and location, see table 1 for office overview of the office characteristics.

Upon examination of the key parameter, WAA duration, the distribution showed that the most common benefit duration among service users across the different NAV-offices was “two years”, with an average of 23.3% of the service users concluding their WAA benefit period during the second year of the trajectory. Conversely, as expected, the “> five years” category had the fewest number of service users. Notably, it is worth mentioning that 13.7% of the service users ended their WAA spell after the standard four-year maximum duration. As shown in Table 1, the median duration ranged from 2.66 to 3.11 years, and the interquartile range (IQR) remained consistent for all NAV offices (IQR = 2).

While there were commonalities in benefit duration patterns, with the majority of cases concentrated around the median, inter-office disparities were evident (see table 2). Notably, office G distinguished itself by exhibiting a higher percentage of cases in the “one year” category (24.5%) and a relatively lower percentage of service users ending their WAA spell within the four- (18.0%) and > five-year (10.8%) categories, compared to the remaining offices. Furthermore, office G also had a slightly higher proportion of its service users falling into the oldest age categories (27.6% and 16.9%) than the average across NAV offices (25.8% and 13.3%). As shown in table 3, service users within the two oldest age categories (50-59 and 60-67) had a higher percentage service users exiting the benefit within the “one year” category (26.0% and 20.7%).

Conversely, we observed that office C exhibits the lowest percentage of cases concluding their benefit period within the first year (14.5%), while also having the highest percentage of cases in the “> five year“ category (19.7%). Examining the demographics at office C, we noted that it served a slightly higher proportion of more female service users (63.5% compared to the NAV office average of 59.0%). Furthermore, we also found that the office C served a greater percentage of unemployed service users (39.7% compared to the general average 34.5%).

We further discovered a significant association between number of activation programs granted and WAA duration (p < 0.001). Upon closer examination of the distribution of activation programs in Table 3, we observed an increasing trend in the granting of activation programs, with the highest number typically allocated in the fifth year. On average, users were allocated 0.61 activation programs in the first year (IQR=1), while users in their fifth year on WAA benefit on average receive 1.31 activation programs (IQR 2). Moreover, significant inter-office variations were found in the granting of activation programs across the local NAV offices (p < 0.001), see Table 2. Office A stood out for providing the highest average number of activation programs (1.06) to the service users. However, office A also exhibited the highest standard deviation, indicating a relatively wide range in the NAV office practice and distribution of activation programs (SD= 1.22). Conversely, office F showed the lowest mean, with an average of 0.66 activation programs granted per service user, and displayed the lowest standard deviation, indicating a consistent practice related to granting of activation programs (SD= 1.03).

This study revealed significant variations in Work Assessment Allowance (WAA) duration among NAV offices operating within the same labour and housing market context. These variations persisted when adjusting for a range of relevant individual variables, indicating a significant association between NAV-office affiliation and service users WAA benefit duration. Particularly, among the seven NAV offices studied, both the largest (office A) and smallest (office C) offices demonstrated statistically significant associations with WAA duration in the adjusted model, holding all other variables at their means. Office C showed the highest average marginal effects (AME), with an average of 4.10 month longer WAA duration (p < 0.001) compared to the medium-sized reference office (F). Meanwhile, office A showed an AME of 1.52 months longer WAA duration (p < 0.05), compared to office F.

The substantial inter-office variations in WAA duration among highly comparable NAV offices, which persist when adjusting for individual variables, suggest the presence of unwarranted variation in service users WAA duration. While our aim was to explore potential variations in benefit duration with the local NAV office as our key predictor, it was not possible to disentangle the potential influence of office characteristics, such as office size, caseload, and centrality of location from the NAV office itself in our regression model. However, we did observe that the size of the local NAV-office appears to be an important factor in our results.

A notable proportion of service users exceeded the standard maximum duration of four years (13.7%). This finding aligns with previous research that has noted prolonged benefit periods in the WAA program (Mandal et al., 2015; Proba, 2015; Sørbø & Ytterborg, 2015). Additionally, evaluations of the WAA program have indicated that inadequate follow-up and capacity issues at NAV offices have led to service users’ work capacity not being determined within the standard four year benefit period (Mandal et al., 2015; Proba, 2015).

Internationally, service user’s experiences of workers compensation claims process have been found to predict benefit duration (Collie et al., 2019), and factors such as prolonged case processing time in workers compensation claims have been associated with negative perceptions of the RTW-process and lower odds of RTW for service users (Cocker et al., 2018; Collie et al., 2019; Sinnott, 2009). Interestingly, the largest NAV office (office A) and the smallest office (office C) exhibited the highest proportion of service users ending their WAA spell in the four- year and >five-year categories.

Previous research within the Norwegian context has identified characteristics related to the size of the local NAV office as important for service provision (Fossestøl et al., 2016; Langeland & Galaasen, 2014; Proba, 2015; Roaldsnes, 2018). Larger NAV offices have been associated with greater caseloads, larger administrative burdens, and more organization and workflow challenges (Fossestøl et al., 2016; Langeland & Galaasen, 2014). Smaller NAV-offices are found to be less susceptible to challenges associated with large caseloads. However, advisors in smaller offices often serve as generalists, providing guidance to service users across multiple benefit areas. These generalist advisors typically face difficulties accessing professional environments and may also have a limited expertise regarding the legal regulations of the WAA program compared to specialised advisors employed in larger NAV-offices (Proba, 2015; Prop. 74L, 2016).

Employees in NAV typically hold a three-year higher education degree as a minimum requirement for employment (Sadeghi & Fekjær, 2019). While about one-third of NAV’s workforce has a background in social work, it is also common to have backgrounds in e.g., economics, administration, and law (Sadeghi & Fekjær, 2019). NAV advisors’ educational background can play an important role in shaping perspectives on work capacity and advisors alignment with policy goals related to labour market orientation (Sadeghi & Fekjær, 2019).

The lack of a uniform educational background in NAV may contribute to unwarranted variations in advisor’s practices and service delivery as the absence of a shared knowledgebase can lead to diverse interpretations and implementation of public service delivery (Van Berkel & Van Der Aa, 2012). For many conditions there exist conflicting or insufficient evidence regarding which measures that best promote RTW. This underscores the importance of advisor’s perspectives and beliefs in guiding their interactions with service users and other stakeholders in the RTW process. Non-specific musculoskeletal disorders represent a case where the absence of clear evidence can impact follow-up efforts. Løchting (2020) found that NAV advisors experienced increased complexity and challenges in facilitating RTW for individuals sick-listed due to non-specific musculoskeletal disorders.

NAV advisors’ perceptions of users’ work capacity and their beliefs about what promotes good health for service users facing health challenges, might also influence activation timing and practices, including advisors’ selection of service users for activation. Previous studies within the Scandinavian context have indicated that the practices of social insurance offices regarding the selection of users for interventions (Ahlgren A. et al., 2008), the timing of interventions, and the use of different types of interventions (Markussen & Røed, 2014) can influence users’ outcomes. Aligning with previous research from the Swedish sickness insurance context (Ahlgren A. et al., 2008; Ahlgren et al., 2005; Marklund et al., 2015), our study observed that the majority of activation programs were allocated during the later stages of the WAA trajectory. Additionally, that the largest NAV office, office A, on average consistently granted a significantly higher number of activation programs and experienced longer average WAA spells. While, office F, on average allocating the fewest activation programs, had notably shorter WAA spells.

In the Norwegian context, delayed implementation of activation measures have commonly been attributed to resource constraints and large caseloads (Proba, 2015; Prop. 74L, 2016). However, in their study of individualization of activation services in NAV, Terum and Sadeghi (2021) surprisingly found no direct relationship between office size and NAV advisors’ reported individualization of activation services. The study underscores that caseload alone may not fully reflect advisor’s workload and that workload perception is influenced by individual preferences and work approaches, impacting employees’ experiences and interactions with service users.

The observed variations in WAA duration and activation practices in our study, emphasizes the importance of further exploring both the effectiveness of activation programs on service users’ outcomes and the contextual factors influencing their implementation. In our analysis, the goodness of fit (R-squared, R2 (in %) = 7.36) indicates relatively low explanatory power. It’s important to acknowledge that the duration of WAA benefit and NAV advisors’ selection to the activation programs is likely influenced by numerous factors not considered in our study. Nevertheless, we observed substantial variations in benefit duration among WAA recipients across NAV offices operating within a shared labor and housing market.

Uncovering unwarranted inter-office variations represents an important opportunity for improving the quality and efficiency of service delivery, especially since practices at the local level can be relatively easily adapted and improved when knowledge is provided on factors contributing to extended benefit durations (Collie et al., 2016). For example, outcome monitoring in mental health services in England revealed substantial between-service variation, with organizational features predicting these outcomes, and improving these organizational aspects led to enhanced outcomes (Clark et al., 2018).

The results of this and similar studies can have direct relevance for practices both within and outside the Nord-Jæren region, and potentially for similar social insurance contexts, such as Sweden and Denmark. By identifying office characteristics or organizational practices associated with longer WAA duration, this research has the potential to inform policy and practice adjustments and potentially ensure more efficient allocation of resources. Additionally, investigating the impact of NAV office size on service delivery, both locally and nationally, holds potential for further exploration. Notably, our study found that the largest and smallest NAV offices exhibited significantly longer WAA durations compared to the medium-sized reference category. Further studies are needed to elaborate the drivers of these variations.

We found significant and substantial variations in service user’s WAA duration across seven local NAV offices within the same labor and housing market context. These variations persisted even after adjusting for a range of relevant individual-level variables. While residual confounding cannot be ruled out, this finding suggests that affiliation with a particular NAV office significantly influences the duration of service user’s WAA benefit. Importantly, our study finds that both the largest and smallest NAV offices exhibit longer WAA durations compared to medium-sized offices, pointing to potential inefficiencies in service delivery and unwarranted variations in service users’ outcomes.

These findings may be valuable for similar administrative contexts across other regions or countries, underscoring the need for further investigations into the underlying mechanisms driving these differences in WAA duration. By understanding and addressing unwarranted variations, we can inform relevant policy and practice adjustments that promote efficiency, quality and equity of public service delivery.

No conflict of interest.