The rare disease neurofibromatosis 1 as a source of hereditary economic inequality : evidence from Finland

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC-ND 4.0 https://creativecommons.org/licenses/by-nc-nd/4.0/ The rare disease neurofibromatosis 1 as a source of hereditary economic inequality : evidence from Finland © 2021 American College of Medical Genetics and Genomics. Published by Elsevier Inc. All rights reserved. Accepted version (Final draft) Johansson, Edvard; Kallionpää, Roope A.; Böckerman, Petri; Peltonen, Sirkku; Peltonen, Juha Johansson, E., Kallionpää, R. A., Böckerman, P., Peltonen, S., & Peltonen, J. (2022). The rare disease neurofibromatosis 1 as a source of hereditary economic inequality : evidence from Finland. Genetics in Medicine, 24(4), 870-879. https://doi.org/10.1016/j.gim.2021.11.024 2022 1 The rare disease neurofibromatosis 1 as a source of hereditary economic inequality: evidence from Finland Edvard Johansson, PhDa,*, Roope A. Kallionpää, MSc (pharm)b,*, Petri Böckerman, PhDc,d,e, Sirkku Peltonen, MD, PhDf,g,h,i, Juha Peltonen, MD, PhDb a Faculty of Social Sciences, Business, and Economics, Åbo Akademi University, Tuomiokirkontori 3, FI- 20500, Turku, Finland b Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FI-20520, Turku, Finland c Jyväskylä University School of Business and Economics, P.O. Box 35, FI-40014, Jyväskylä, Finland d Labour Institute for Economic Research, Helsinki, Finland e IZA Institute of Labor Economics, Bonn, Germany f Department of Dermatology and Venereology, University of Turku, P.O. Box 52, 20521, Turku, Finland g Department of Dermatology, Turku University Hospital, Turku, Finland h Department of Dermatology and Venereology, University of Gothenburg, Gothenburg, Sweden i Department of Dermatology and Venereology, Sahlgrenska University Hospital, Gothenburg, Sweden * These authors contributed equally to the study. Corresponding author: Juha Peltonen, Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FI-20520 Turku, Finland; tel. +358 29 450 4613; ju[email protected]i; fax +358 29 450 5040. Running title: Rare disease and economic inequality 2 Conflict of Interest The authors declare no conflict of interest. 3 Abstract Purpose: This study investigates whether individuals with neurofibromatosis 1 (NF1) fare worse than individuals without NF1 in terms of economic well-being. NF1 is relatively common in the population and provides an informative case of a rare hereditary disease. Methods: We examined a subset of 692 individuals with verified NF1 from the Finnish total populationbased NF1 cohort, and compared that to 7,407 control individuals matched for age, sex and municipality during 1997-2014. Economic well-being was operationalized with annual work earnings and total income including social income transfers. Results: NF1 significantly worsened economic well-being. Low education, increased morbidity, and reduced labor market participation partly explained the effect of NF1. Yet, NF1 was independently associated with lower income even after adjusting for these factors. Further, NF1 had a larger negative effect on income from work than it had on total income, which indicates that the Finnish social security system partly compensated the labor market losses suffered by individuals with NF1. NF1 had a larger impact on economic inequality for men than women. Conclusions: NF1 contributes to economic inequality. A hereditary disease may convey worse economic well-being over several generations. Keywords Neurofibromatosis, Nordic model, rare diseases, social income transfers, wages 4 Introduction Although each rare disease has an incidence of less than 1/2,000, rare diseases affect, in total, 4-6% of the population globally.1 It has been estimated that there are 27-36 million individuals with a rare disease in the European Union, and 25-30 million in the US.2 Most rare diseases are heritable, life threatening, or chronically debilitating complex diseases. Although rare diseases vary in their clinical characteristics, patients share many experiences, such as difficulties in getting a diagnosis and in obtaining information about their disease, paucity of treatment options, poor resources for rehabilitation, and lack of peer support.2 Although these shortcomings in health care services have been recognized, only little empirical research has been conducted on the effect of rare diseases on economic well-being and life in general. Carrying out large register-based studies on rare diseases is challenging, since collecting large datasets is impossible in the case of most rare diseases. The Finnish nationwide registers provide an excellent source for studying rare diseases. Neurofibromatosis 1 (NF1; OMIM 162200) was chosen as a model disease for this multidisciplinary study, as NF1 is sufficiently common to allow a case-control cohort study. NF1 affects about 1 in 2,000-3,000 persons worldwide.3,4 NF1 is a congenital multiorgan syndrome caused by pathogenic variants of the NF1 gene.5,6 NF1 has variable cutaneous, neural, and skeletal manifestations which may be present from infancy.7 NF1 is associated with a 60% lifetime risk of cancer8 and with significant excess mortality in all age groups.3,4 Learning disabilities and cognitive and behavioral disorders are also common among individuals with NF1. A recent study found that NF1 is associated with reduced educational attainment, and the affected individuals have a tendency for vocational rather than academic educations.9 Despite its specific characteristics, NF1 may point to aspects that need to be considered in the context of other rare diseases. For example, the tuberous sclerosis complex is a genetic disorder characterized by potentially disfiguring cutaneous manifestations, learning disabilities, and various tumors;10 and the Lynch syndrome is associated with a similar level of cancer risk as NF1.11 Like many other rare diseases, NF1 is dominantly inherited, and approximately half of the individuals with NF1 have a parent with NF1. A sibling or a parent with NF1 may reduce a healthy family member’s probability of obtaining academic education.9 Moreover, the cognitive development and intelligence 5 quotient (IQ) of children with familial NF1 may be inferior to those of individuals with sporadic disease.12,13 It is therefore important to examine whether familial NF1 contributes to long-term labor market success. The cost of illness of rare diseases has been studied from a societal perspective.14 The effects of cancer and health shocks or the effects of disability on labor market performance and earnings in general have also been studied.15–17 However, studies concerning the indirect individual costs in terms of reduced labor market performance or earnings of individuals with rare heritable diseases are scarce, or non-existent. This is also the case for NF1. In order to better understand the effects of a rare, inherited disease on long-term labor market success and income inequality, we used the total population-based Finnish NF1 cohort and several nationwide registers to elucidate how NF1 affects labor market performance of individuals with familial or sporadic NF1 in the context of the Nordic model. By the Nordic model we refer to the societal model of the Nordic countries.18 These countries share many characteristics, the most important of which in this context is the relatively high level of taxes and social income transfers. The aim of these policies is to moderate differences in disposable income. 6 Materials and methods The study was approved by the Ethics committee of the Hospital District of Southwest Finland and research permissions were obtained from the Finnish Institute for Health and Welfare, Statistics Finland, and all participating hospitals. The study adhered to the principles set out in the Declaration of Helsinki. Individuals with NF1 were compared to control individuals over a study period of 1997-2014. Individuals fulfilling the National Institutes of Health (NIH) diagnostic criteria for NF1 were identified by searching the 5 University Hospitals and 15 Central Hospitals of mainland Finland for NF1-associated hospital visits in 1987-2011. For this, the International Classification of Diseases (ICD), 9th edition code 2377A and ICD-10 codes Q85.00, Q85.0, Q85.09, Q85, and Q85.01 were used.3 The medical records of each patient were reviewed to confirm that the NIH diagnostic criteria were fulfilled. The first NF1-related hospital visit was considered as the cohort entry. All analyses were restricted to persons aged 25-64, since most individuals within this age range have completed their studies but have not retired. Moreover, at least five years of data were required for each individual. A total of 692 individuals with NF1 fulfilling these criteria were identified, representing 594 different families. The NF1 syndrome was considered familial if a parent with confirmed NF1 was known, if familial disease was documented in the medical records, or if the individual with NF1 had at least one sibling with NF1. For each individual in the full Finnish NF1 cohort, ten control individuals without NF1 matched for age, sex and municipality were retrieved from the Finnish Population Register Centre. The Finnish Population Register Centre keeps track of all inhabitants in Finland and records such data as date of birth, death and emigration, and all family relationships, e.g., parents, children, and siblings. By matching the controls with the NF1 cohort, the risk of bias related to temporal trends or the area of residence, such as the distance to health care service providers or educational institutions, was reduced. The cohort entry date of the respective individual with NF1 was used as the start of follow-up for the controls. First-degree relatives of 7 individuals with NF1 were excluded from the control cohort. Consequently, none of the control individuals had a parent with known NF1. The final number of individuals in the control cohort was 7,407. Five outcomes were studied: 1) The dichotomous indicator whether an individual earned any income from work during a year or not. 2) Among individuals who had income from work, the amount of income earned per year. Work income was deflated using the consumer price index provided by Statistics Finland to account for inflation and to allow comparability over the 18-year study period, and natural logarithm transformation was used to facilitate interpretation of the estimates. 3) The dichotomous indicator whether an individual had received social income transfers from the public sector or not. All transfers from the public sector during a year were included, except parental benefits. The transfers included items like pension income, unemployment benefits, study grants, sickness allowance and social assistance. 4) Among individuals with positive social income transfers, the amount of transfers per year. The amount of transfers was deflated using the consumer price index, and natural logarithm transformation was used. 5) The natural logarithm of the amount of total income per year, defined as the sum of the deflated work income and social income transfers. The Finnish personal identity code was used as the key when information on work income, social income transfers, working months, days on sick leave and educational attainment were retrieved from Statistics Finland. The Finnish personal identity code is an immutable identifier assigned to each person at birth, and it is used to record information in the national population-based registries and for example in health care, thus allowing longitudinal follow-up with essentially no data loss. Hospital visits and hospital stays were obtained from the Finnish Care Register for Health Care which records all inpatient care and specialized outpatient care. 8 In all statistical models, individuals with and without NF1 were compared. All analyses were adjusted for sex, age, and the square of age to account for both linear and non-linear effects of increasing age.19 In addition, models were constructed with adjustments for educational attainment, the number of hospital visits and hospital stays, as well as having inherited NF1, i.e., NF1 being familial or sporadic. The number of working months during the year was included in the analyses of annual earnings, amount of social income transfers, and total income. In the analyses of annual earnings and total income, the number of days on sick leave during the year was also accounted for. In terms of educational attainment, secondary education (International Standard Classification of Education, ISCED 3-5) and tertiary education (ISCED ≥6) were included as separate variables. The number of hospital visits and hospital stays during the year was included to account for NF1-related morbidity. The follow-up of each individual ended at death, emigration, or the end of follow-up period in 2014, and the analyses always compare the individuals surviving to each age. The analyses were performed using linear panel data regression analysis which allowed quantification of the effect of NF1 on economic outcomes. The use of regression analysis makes it also possible to examine the contributions of education, labor market participation, and other pertinent confounders for explaining the association between NF1 and economic outcomes. Random effects were used to group the observations from different years of each individual. Thus, an observation is a person-year, and time (yearly) effects were controlled for. In all models, standard errors were further clustered within the strata of one individual with NF1 and the maximum of 10 matched control individuals to account for the matching of controls to individuals with NF1. The statistical analysis was conducted using Stata software version 15. 15 Data Availability Data are available upon request for researchers though data access is restricted. Please contact the Finnish National Institute for Health and Welfare and Statistics Finland for permission. Data can be requested from the corresponding author Prof. Juha Peltonen: Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FI-20520 Turku, Finland; [email protected]. Acknowledgements The study was funded with grants from the Turku University Hospital and the Cancer Foundation Finland. This work is generated within the European Reference Network on Genetic Tumour Risk Syndromes (ERN GENTURIS)—Project ID No 739547. ERN GENTURIS is partly cofunded by the European Union within the framework of the Third Health Programme “ERN- 2016—Framework Partnership Agreement 2017–2021”. Author Information Conceptualization: all authors; Data curation: EJ, RAK, SP, JP; Formal Analysis: EJ; Funding acquisition: SP, JP; Investigation: all authors; Writing: all authors. Ethics Declaration The study was approved by the Ethics committee of the Hospital District of Southwest Finland and research permissions were secured from the Finnish Institute for Health and Welfare, Statistics Finland, and all participating hospitals. The study adhered to the principles set out in the Declaration of Helsinki. The study is register-based and retrospective and therefore exempt from obtaining informed consent from the participants. 16 References 1. Nguengang Wakap S, Lambert DM, Olry A, et al. Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database. Eur J Hum Genet. 2020;28(2):165-173. doi:10.1038/s41431-019-0508-0 2. De Vrueh R, Baekelandt ERF, De Haan JMH. Background paper 6.19 rare diseases. World Heal Organ Geneva. Published online 2013. 3. Uusitalo E, Leppävirta J, Koffert A, et al. Incidence and mortality of neurofibromatosis: a total population study in Finland. J Invest Dermatol. 2015;135(3):904-906. doi:10.1038/jid.2014.465 4. Kallionpää RA, Uusitalo E, Leppävirta J, Pöyhönen M, Peltonen S, Peltonen J. Prevalence of neurofibromatosis type 1 in the Finnish population. Genet Med. 2018;20(9):1082-1086. doi:10.1038/gim.2017.215 5. Wallace MR, Marchuk DA, Andersen LB, et al. Type 1 neurofibromatosis gene: Identification of a large transcript disrupted in three NF1 patients. Science. 1990;249(4965):181-186. doi:10.1126/science.2134734 6. Xu G, O’Connell P, Viskochil D, et al. The neurofibromatosis type 1 gene encodes a protein related to GAP. Cell. 1990;62(3):599-608. doi:10.1016/0092-8674(90)90024-9 7. Legius E, Messiaen L, Wolkenstein P, et al. Revised diagnostic criteria for neurofibromatosis type 1 and Legius syndrome: an international consensus recommendation. Genet Med. 2021;23(8):1506-1513. doi:10.1038/s41436-021-01170-5 8. Uusitalo E, Rantanen M, Kallionpää RA, et al. Distinctive Cancer Associations in Patients With Neurofibromatosis Type 1. J Clin Oncol. 2016;34(17):1978-1986. doi:10.1200/JCO.2015.65.3576 9. Johansson E, Kallionpää RA, Böckerman P, Peltonen J, Peltonen S. A rare disease and education: Neurofibromatosis type 1 decreases educational attainment. Clin Genet. 2021;99(4):529-539. doi:10.1111/cge.13907 17 10. Leung AKC, Robson WLM. Tuberous sclerosis complex: a review. J Pediatr Health Care. 21(2):108-114. doi:10.1016/j.pedhc.2006.05.004 11. Bucksch K, Zachariae S, Aretz S, et al. Cancer risks in Lynch syndrome, Lynch-like syndrome, and familial colorectal cancer type X: a prospective cohort study. BMC Cancer. 2020;20(1):460. doi:10.1186/s12885-020-06926-x 12. Biotteau M, Déjean S, Lelong S, et al. Sporadic and Familial Variants in NF1: An Explanation of the Wide Variability in Neurocognitive Phenotype? Front Neurol. 2020;11:368. doi:10.3389/fneur.2020.00368 13. Hou Y, Allen T, Wolters PL, et al. Predictors of cognitive development in children with neurofibromatosis type 1 and plexiform neurofibromas. Dev Med Child Neurol. 2020;62(8):977- 984. doi:10.1111/dmcn.14489 14. Angelis A, Tordrup D, Kanavos P. Socio-economic burden of rare diseases: A systematic review of cost of illness evidence. Health Policy. 2015;119(7):964-979. doi:10.1016/j.healthpol.2014.12.016 15. Currie J, Madrian B. Health, health insurance and the labor market. In: Ashenfelter O, Card D, eds. Vol 3, Part C. Elsevier; 1999:3309-3416 BT-Handbook of Labor Economics. https://econpapers.repec.org/RePEc:eee:labchp:3-50 16. Vaalavuo M. The unequal impact of ill health: Earnings, employment, and mental health among breast cancer survivors in Finland. Labour Econ. 2021;69:101967. doi:https://doi.org/10.1016/j.labeco.2021.101967 17. Jones M. Disability and labor market outcomes. IZA World Labor. Published online 2021. 18. Andersen TM, Holmström B, Honkapohja S, Korkman S, Söderström HT, Vartiainen J. The Nordic Model. Embracing Globalization and Sharing Risks. ETLA B232. The Research Institute of the Finnish Economy; 2007. 19. Rosenzweig MR. Nonlinear earnings functions, age, and experience: a nondogmatic reply and 18 some additional evidence. J Hum Resour. 1976;11(1):23-27. 20. Currie J. Healthy, Wealthy, and Wise: Socioeconomic Status, Poor Health in Childhood, and Human Capital Development. J Econ Lit. 2009;47(1):87-122. http://www.jstor.org/stable/27647135 21. Ettner SL. New evidence on the relationship between income and health. J Health Econ. 1996;15(1):67-85. doi:https://doi.org/10.1016/0167-6296(95)00032-1 22. Viinikainen J, Bryson A, Böckerman P, et al. Do childhood infections affect labour market outcomes in adulthood and, if so, how? Econ Hum Biol. 2020;37:100857. doi:https://doi.org/10.1016/j.ehb.2020.100857 23. Doser K, Kenborg L, Andersen EW, et al. Educational delay and attainment in persons with neurofibromatosis 1 in Denmark. Eur J Hum Genet. 2019;27(6):857-868. doi:10.1038/s41431-019- 0359-8 24. Vogel AC, Gutmann DH, Morris SM. Neurodevelopmental disorders in children with neurofibromatosis type 1. Dev Med Child Neurol. 2017;59(11):1112-1116. doi:10.1111/dmcn.13526 25. Krab LC, Aarsen FK, de Goede-Bolder A, et al. Impact of Neurofibromatosis Type 1 on School Performance. J Child Neurol. 2008;23(9):1002-1010. doi:10.1177/0883073808316366 26. Plasschaert E, Van Eylen L, Descheemaeker MJ, Noens I, Legius E, Steyaert J. Executive functioning deficits in children with neurofibromatosis type 1: The influence of intellectual and social functioning. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(3):348-362. doi:10.1002/ajmg.b.32414 27. Hyman SL, Arthur E, North KN. Learning disabilities in children with neurofibromatosis type 1: subtypes, cognitive profile, and attention-deficit- hyperactivity disorder. Dev Med Child Neurol. 2006;48(12):973-977. doi:10.1111/j.1469-8749.2006.tb01268.x 28. Mautner VF, Granström S, Leark RA. Impact of ADHD in Adults With Neurofibromatosis Type 1: Associated Psychological and Social Problems. J Atten Disord. 2015;19(1):35-43. 19 doi:10.1177/1087054712450749 29. Ferner RE, Hughes RAC, Weinman J. Intellectual impairment in neurofibromatosis 1. J Neurol Sci. 1996;138(1-2):125-133. doi:10.1016/0022-510X(96)00022-6 30. Ameri M, Schur L, Adya M, Bentley FS, McKay P, Kruse D. The Disability Employment Puzzle: A Field Experiment on Employer Hiring Behavior. ILR Rev. 2018;71(2):329-364. doi:10.1177/0019793917717474 20 Figure legends Figure 1. Work income (A), social income transfers (B) and total income (C) by age group among individuals with NF1 and controls.     ,QFRPHSHU\HDU¼ :RUNLQFRPH     7UDQVIHUVSHU\HDU¼ 6RFLDOLQFRPHWUDQVIHUV      7RWDOLQFRPHSHU\HDU¼  ± ± ± ± ± ±  ± ± 7RWDOLQFRPH &RQWUROV ,QGLYLGXDOVZLWK1) $JH\HDUV  ± ± ± ± ± ±  ± ± $JH\HDUV  ± ± ± ± ± ±  ± ± $JH\HDUV $ % & )LJXUH Table 1. Characteristics of the NF1 and control cohorts. Sporadic NF1 Familial NF1 All NF1 Controls N 418 274 692 7,407 Age (years), mean (SD) 43.99 40.00b 42.44 42.59 (10.90) (10.07) (10.76) (10.87) Age at cohort entry (years), mean (SD) 34.23 29.42b 32.36 32.02a (13.31) (13.72) (13.67) (13.71) Years in sample, mean (SD) 15.71 15.37 15.79 15.58 (3.55) (3.75) (3.51) (3.63) Females, % (N) 56 57 56 56 (234) (156) (390) (4,148) Months worked /year, mean (SD) 7.21 7.58b 7.35 8.92b (5.47) (5.35) (5.43) (4.82) Number of hospital visits /year, mean (SD) 2.39 2.75b 2.53 1.33b (5.06) (6.23) (5.55) (4.52) Number of sick days /year, mean (SD) 6.96 6.99 6.97 4.72b (30.38) (31.31) (30.74) (23.86) Any income earned, % (N) 70 73b 71 84b (292) (199) (491) (6,222) Average annual work income, mean (SD) 17,127 17,115 17,123 26,151b (13,749) (13,568) (13,678) (23,197) Transfers received, % (N) 67 71b 69 58b (280) (195) (475) (4,296) Average annual transfers, mean (SD) 6,235 5,891a 6,102 4,336b (5,748) (5,924) (5,819) (4,928) Average annual total income, mean (SD) 23,375 23,007 23,232 30,491b (10,688) (10,902) (10,772) (21,636) Less than secondary education, % (N) 22 23 22 19b (92) (62) (154) (1,407) Secondary education, % (N) 57 63b 59 44b (237) (173) (410) (3,259) Tertiary education, % (N) 21 14b 18 37b (88) (38) (126) (2,741) Observations 5,996 3,807 9,803 107,147 a denotes differences in means that are statistically significant at the 5% level; b denotes differences in means that are statistically significant at 1% level;. Footnotes in column 2 indicate differences between sporadic NF1 and familial NF1. Footnotes in column 4 indicate differences between All NF1 and Controls. Table 2: The effect of NF1 on labor market outcomes (1) Probability of earning work income (2) Log of annual earnings (3) Probability of obtaining positive amount of social income transfers (4) Log of social income transfers (5) Log of total income Individual has NF1 -0.131c -0.119c -0.305c -0.111b 0.113c 0.064c 0.336c 0.104b -0.199c -0.047a (0.013) (0.016) (0.044) (0.037) (0.012) (0.016) (0.037) (0.033) (0.020) (0.019) NF1 x inherited NF1 0.022 -0.040 0.026 0.034 0.017 (0.025) (0.056) (0.024) (0.055) (0.031) Number of hospital visits / year -0.003c -0.007c 0.007c 0.001 -0.000 (0.000) (0.001) (0.001) (0.001) (0.000) Number of months working / year 0.170c -0.126c 0.057c (0.002) (0.002) (0.001) Number of sick days / year -0.001c 0.000c (0.000) (0.000) Secondary education 0.155c 0.169c -0.050c -0.051 0.090c (0.011) (0.021) (0.010) (0.028) (0.013) Tertiary education 0.206c 0.548c -0.200c -0.076a 0.432c (0.011) (0.025) (0.012) (0.033) (0.016) Age 0.041c 0.039c 0.184c 0.078c -0.028c -0.025c -0.119c 0.011a 0.112c 0.053c (0.002) (0.002) (0.006) (0.004) (0.003) (0.003) (0.008) (0.006) (0.003) (0.002) Age squared -0.001c -0.001c -0.002c -0.001c 0.000c 0.000c 0.002c 0.000 -0.001c -0.000c (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.016a -0.024c -0.336c -0.327c 0.256c 0.269c 0.222c 0.289c -0.221c -0.238c (0.008) (0.007) (0.031) (0.018) (0.010) (0.010) (0.032) (0.023) (0.021) (0.020) N 116,950 116,950 96,886 96,886 116,950 116,950 68,502 68,502 115,252 115,252 Note: The data are model estimates with standard errors in parentheses. Observations are clustered within strata of 1 individual with NF1 and the maximum of 10 control individuals without NF1. The footnotes denote statistical significance: a: 0.05 > P ≥ 0.01; b: 0.01 > P ≥ 0.001; c: P < 0.001 Table 3: The effect of NF1 on labor market outcomes: women (1) Probability of earning work income (2) Log of annual earnings (3) Probability of obtaining positive amount of social income transfers (4) Log of social income transfers (5) Log of total income Individual has NF1 -0.134c -0.116c -0.258c -0.023 0.069c 0.012 0.241c 0.028 -0.174c -0.020 (0.019) (0.022) (0.059) (0.048) (0.016) (0.020) (0.044) (0.040) (0.028) (0.026) NF1 x inherited NF1 0.013 -0.154 0.075a 0.076 -0.001 (0.036) (0.081) (0.029) (0.062) (0.042) Number of hospital visits / year -0.003c -0.008c 0.005c -0.002 -0.002c (0.000) (0.001) (0.001) (0.001) (0.001) Number of months working / year 0.171c -0.112c 0.055c (0.003) (0.002) (0.001) Number of sick days / year -0.000 0.002c 0.000c (0.000) (0.000) (0.000) Secondary education 0.168c 0.157c -0.042b -0.075a 0.068c (0.015) (0.031) (0.013) (0.032) (0.019) Tertiary education 0.222c 0.557c -0.139c -0.068a 0.388c (0.016) (0.035) (0.014) (0.033) (0.022) Age 0.049c 0.046c 0.186c 0.068c 0.000 0.003 -0.119c 0.015a 0.115c 0.052c (0.003) (0.003) (0.008) (0.006) (0.004) (0.005) (0.007) (0.007) (0.004) (0.003) Age squared -0.001c -0.001c -0.002c -0.001c -0.000 -0.000 0.002c -0.000 -0.001c -0.000c (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) N 65,909 65,909 54,173 54,173 65,909 65,909 46,598 46,598 65,045 65,045 Note: The data are model estimates with standard errors in parentheses. Observations are clustered within strata of 1 individual with NF1 and the maximum of 10 control individuals without NF1. The footnotes denote statistical significance: a: 0.05 > P ≥ 0.01; b: 0.01 > P ≥ 0.001; c: P < 0.001