Employment, occupation, and income in adults with neurofibromatosis 1 in Denmark: a population- and register-based cohort study

We have previously established a national cohort of 2467 individuals hospitalized with or for NF1 or followed on one of two national Centers for Rare Diseases at Aarhus University Hospital or Copenhagen University Hospital, Rigshospitalet (a detailed description of the cohort is available [16]) in the period 1977 to 2013. All individuals with NF1 were matched to ten general population comparisons, randomly selected from the Danish Civil Registration System and matched on birth month and year as well as sex [17]. From this cohort, we identified 1469 individuals with NF1 living in Denmark and their 11,991 matched population comparisons, who were ≥ 30 years old at start of study. Finally, we retrieved information on migration and vital status from the Danish Civil Registration System for all cohort members.

All inhabitants in Denmark are given a unique personal identification number at birth or upon emigration to Denmark, which enables accurate linkage among national registries [17]. By linking the personal identification number of all included individuals, we ascertained information on employment, occupation and income from nationwide administrative registers updated annually by Statistics Denmark.

Information on annual employment and disposable income was obtained from the Integrated Database for Labour Market Research (IDA Register) [18], which was available in the period 1980–2019. The register contains longitudinal information on income, employment, and establishments on an individual level. We categorized employment status on annual basis into three groups: employed, unemployed (i.e., individuals who receive unemployment benefits and not health-related social security benefits during that calendar year) and unemployed for health-related reasons (i.e., individuals who received sickness or disability benefits, social assistance or rehabilitation benefits or health-related retirement during that calendar year). If an individual was registered as employed at one point during a calendar year, then the individual was categorized as employed irrespective of any other registration of unemployment during the same calendar year. Likewise, an individual was classified as unemployed for health-related reasons if any social benefits due to health reasons were received during a calendar year. Thus, employment, unemployment, and health-related unemployment were mutually exclusive, so an individual could only be registered as either employed, unemployed, or unemployed for health-related reasons during a calendar year. Finally, we did not include individuals in the employment analyses in a calendar year during which they were enrolled as students at any teaching institutions.

Also, from the IDA Register [18], we included data on annual disposable income, which is defined as the amount of money that an individual has in his or her possession after income taxes and interests have been deducted. Based on annual disposable income, we defined low income as income below the threshold set at 60% of the median disposable income, corresponding to the at-risk-of-poverty threshold used by Eurostat [19].

We obtained information on occupational positions from the Employment Classification Module (AKM Register), which collects annual information on the most important employment activity (i.e. the activity with the highest income) since 1993 [18]. We grouped occupational positions for the period 1993–2019 into 10 major groups according to the International Standard Classification of Occupation (ISCO-08) [20]: (1) managers, (2) professionals, (3) technicians and associate professionals, (4) clerical support workers, (5) services and sales workers, (6) skilled agricultural, forestry and fishery workers, (7) craft and related trades workers, (8) plant and machine operators and assemblers, (9) elementary occupations, (10) armed forces occupations. Additionally, we grouped occupational positions according to ISCO skill level (i.e., the ability to carry out the tasks and duties of a given occupation defined by the complexity and range of tasks) with highly skilled occupations defined by ISCO group 1, 2 or 3 [20].

We have previously reported that individuals with NF1 are less likely to achieve a medium or long education compared to the background population [9]. Thus, we linked our study population to the Danish Education Registry, which includes information on all educational programs in Denmark since 1981. The Registry ascertains the highest level of completed education on a yearly basis [21]. We included information on highest attained education before age 30 years categorized as short (≤ 9 years of education), medium (10–12 years), and long (> 12 years) education.

We included information on somatic and psychiatric disease before age 30 years to assess their association with employment status, occupational position, and disposable income at age 30 years. From the Danish Cancer Registry [22], we retrieved all primary cancer diagnoses. Information on hospitalizations for somatic disease other than cancer was obtained from the Danish National Patient Registry [23] and categorized into six groups on a yes/no level: benign tumors, neurologic disease, gastrointestinal disease, diseases of bones, joints and soft tissue and other somatic disease. Finally, we included information on psychiatric hospital contacts from the Danish Psychiatric Central Research Register [24]. The register was established in 1969 and includes information on all patients treated at psychiatric departments in Denmark, supplemented with outpatient and emergency room contacts since 1995. Information on psychiatric diagnoses was included as any/no hospital contacts for psychiatric disease (a list of all included diagnoses has previously been published [5]).

We assessed employment, health-related unemployment, and other unemployment as well as disposable income from age 30 years using the study cohorts of 1469 individuals with NF1 and 11,991 population comparisons. The study population was followed until date of death, emigration, or end of study period (31 December 2019), whichever occurred first. Only individuals surviving to each age were compared.

Employment, health-related unemployment, other employment, and low disposable income were defined as dichotomous outcomes. We plotted the annual prevalence of individuals in each group between ages 30 and 50 years stratified by sex. Unconditional logistic regression models were used to estimate odds ratios and 95% confidence intervals (CIs) for employment, health-related unemployment, other employment, and low disposable income adjusting for calendar period using linear splines, highest attained education, and sex. In all analyses, individuals with and without NF1 were compared at age 30, 40, and 50 years. The analyses were conducted using the whole study population alive at each age. We also conducted logistic regression analyses of employment, unemployment, and health-related unemployment at age 30 years for specific somatic disease groups (cancer, benign tumors, diseases of nervous system and sense organs, diseases of digestive organs, and diseases of bone, joints, and soft tissue) and any psychiatric disease adjusting for calendar period using linear splines, highest attained education, and sex.

We plotted the median annual disposable income between ages 30 to 50 years. We used linear regression to assess the disposable income level in individuals with NF1 compared to the general population for different birth cohorts (1929–1940, 1941–1950, 1951–1960, 1961–1970, 1971–1980, and 1981–1988). We also assessed difference in occupational position at age 30 years using the 10 major ISCO groups in a logistic regression model. Employed cohort members with missing information on occupation were included as a missing group. Finally, we estimated the odds for obtaining a highly skilled occupation (defined as ISCO group 1, 2 and 3) at ages 30, 40 and 50 years using a logistic regression model. These analyses were adjusted for calendar year using linear splines, highest attained education, and sex. We performed the analyses using R version 4.1.0 with the spline packages [25].

Figure 1A–C show the annual proportions of employment, unemployment, and health-related unemployment at ages 30 to 50 years. At all ages, individuals with NF1 were less often employed than the comparisons, with the lowest employment rate observed in women with NF1 aged 50 years (63%; female comparisons: 79%) (Fig. 1A). The annual proportion of individuals being unemployed was similar to the comparison cohort; however, individuals with NF1 were more often unemployed due to reasons related to health. At age 50 years, 18% of women with NF1 were health-related unemployed compared to 7% of the female comparisons.

We found that individuals with NF1 had a lower OR for employment than the comparison cohort, both at age 30 (OR 0.71, 95% CI 0.61–0.83), 40 (OR 0.58, 95% CI 0.49–0.68), and 50 (0.53, 95% CI 0.45–0.62) years (Table 2). We did not find any increased OR for unemployment. At age 30 years, individuals with NF1 had an increased OR for health-related unemployment of 2.94 (95% CI 2.16–3.96), which was also increased at age 40 and 50 years. The ORs for employment were slightly lower in women with NF1 than in men with NF1 when compared with women and men from the general population, respectively. The ORs for health-related unemployment were slightly higher in women with NF1 (Additional file 1: Table S1).

We assessed previous somatic hospitalizations for cancer, benign tumors, neurological disease, gastrointestinal disease and diseases of bones, joints and soft tissues and psychiatric contacts as risk factors for employment status at age 30 years (Table 3). Individuals with NF1 who had somatic hospitalizations before age 30 years had a higher OR for being unemployed for health-related reasons than population comparisons who also had somatic hospitalizations. Due to too few cancer diagnoses in the comparison cohort, we were unable to assess cancer as a risk factor for health-related unemployment. No association was found between any psychiatric hospital contact and the different outcomes.

Individuals with NF1 had lower ORs for being in high skilled occupation, including managers, professionals, and technicians and associate professionals at age 30 years compared with individuals without NF1 (Table 4); however, the OR was only significantly decreased for professionals (OR 0.60, 95% CI 0.41–0.87). When we combined the high skilled occupations, individuals with NF1 had significantly lower ORs for being in a high skilled occupation both at age 30, 40 and 50 years (Table 4).

Figure 2 shows the median, 25th and 75th percentile annual disposable income by age for different birth cohorts. The difference in disposable income between individuals with NF1 and the comparison group increased for the youngest birth cohorts, where the median disposable income for those with NF1 was almost similar to the 25th percentile of the comparison group. The results from the linear regression revealed that adults with NF1 had a slightly lower disposable income than the population comparisons, with a 14% reduction observed among the youngest birth cohort (Additional file 2: Table S2). However, individuals with NF1 did not have increased ORs for being in the low disposable income group either at age 30, 40 or 50 years (Table 2).

The strengths of the study include the population-based design with information on employment, income, and occupation from high-quality nationwide registries, irrespective of NF1 status, which reduced the risk of information bias from self-reported data and limited loss to follow-up. We were also able to compare the employment history in individuals with NF1 with that of randomly selected population comparisons to represent the Danish background population. The NF1 patients identified in the two Centers for Rare Diseases fulfilled the US National Institutes of Health (NIH) criteria for NF1 [32] or had their diagnosis confirmed by molecular genetic testing. For the patients identified in the Danish Patients Register, we excluded those with clinical and anatomic features of NF2, but some misclassification of the NF1 diagnosis cannot be ruled out. As the individuals in the NF1 cohort had all been hospitalized with or for NF1 or are followed in the Centers for Rare Diseases, we might have missed individuals with NF1 who are less affected by their NF1. However, all Danish inhabitants are encouraged to seek follow-up care at one of the two Centers for Rare Diseases, regardless of the severity of their disease. Follow-up services are provided free of charge, which reduces the risk of selection of participants due to financial constraints. In addition, the process of diagnosing NF1 will result in at least one registration in the RAREDIS Database, even if the patient declines further follow-up at the Center for Rare Diseases. Despite these initiatives, some individuals with mild symptoms or asymptomatic relatives may choose not to undergo clinical investigation for NF1, and consequently, they will not be included in our NF1 cohort. The employment rates and income equality might have been improved if we were able to include study participants not affiliated with the two Centers for Rare Diseases. Thus, our results should not be generalized to adults with NF1 who are only in contact with the primary health care system. Furthermore, we included some who received both benefits within the same calendar year in our definition of unemployment and health-related employment. However, this potential misclassification is expected to be similar for adults with and without NF1, thus biasing towards the null. Another limitation of the study is the inclusion of psychiatric hospital contacts using the Danish Psychiatric Central Research Register, which only holds information on outpatient contacts since 1995. Most patients with depression and anxiety as well as attention deficit/hyperactivity disorder and autism spectrum disorders are not hospitalized. Consequently, we were not able to assess the association between specific psychiatric disorders and employment status. Finally, the study was conducted in Denmark, where social security benefits are higher than in other countries. Thus, the gap in disposable income between adults with and without NF1 might be higher in countries outside Denmark and other Nordic countries.