Motor proficiency of persons with attention deficit hyperactivity disorder or autism spectrum disorder diagnosed in adulthood

Abstract

Purpose

To compare (1) motor proficiency of persons diagnosed in adulthood with attention deficit hyperactivity disorder (ADHD) or autism spectrum disorder (ASD) with normative values of motor proficiency, and (2) motor proficiency between persons with ADHD and those with ASD diagnosed in adulthood.

Methods

A total of 153 adults (median age 32 years, 36% women) participated in this cross-sectional study. Fifty-three persons with predominately inattentive presentation (ADHD-I), 67 persons with combined presentation (ADHD-C), and 33 persons with ASD performed the Bruininks-Oseretsky Test of Motor Proficiency (BOT-2). One-sample binominal tests were used to compare motor proficiency against standardized norms of BOT-2 for young adults. One-way ANOVAs and Kruskal-Wallis tests were used to compare test outcomes between the groups.

Results

The total sample showed significantly impaired motor proficiency in comparison to norms in all test domains (p < 0.001–0.006), except for fine motor skills. The ASD group showed significantly poorer body coordination compared with the ADHD-I and ADHD-C groups, with a moderate effect size (p = 0.003–0.02, η² = 0.061).

Conclusions

Motor proficiency is impaired in most persons with ADHD or ASD diagnosed in adulthood, suggesting that motor assessment should be included in clinical examinations of adults with suspected neurodevelopment disorders.

IMPLICATION FOR REHABILITATION

• Motor proficiency is impaired in most adults with adult-diagnosed attention deficit hyperactivity disorder (ADHD) or autism spectrum disorder (ASD).

• Body coordination is more impaired in adults with ASD than adults with ADHD.

• Motor assessment of adults with suspected neurodevelopment disorders is recommended.

Keywords:

• Attention deficit hyperactivity disorder
• autism spectrum disorder
• adults
• Bruininks-Oseretsky test of motor proficiency
• motor skills

Introduction

Attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) are neurodevelopmental disorders with a worldwide prevalence and are among the fastest-growing diagnostic groups in the field of psychiatry [1]. The prevalence of ADHD and ASD in children and adults varies greatly across the world, likely explained by methodological differences and characteristics of studies [2–5]. Among adults, the ADHD prevalence has been averaged to 2.8% with a higher prevalence in high-income countries (3.6%) than in low/lower-middle-income countries (1.4%) [3] whereas the ASD prevalence in high-income countries such as England, Sweden, and the United States has been estimated to 0.9–2.2% [4, 6, 7].

Motor skill impairments are common in children with ADHD or ASD, with a higher prevalence in children with ASD [8]. The prevalence rate of motor impairments has been estimated to be more than 50% and 80% among children with ADHD [9], and ASD [10], respectively. Despite comprehensive research on children with ADHD or ASD, it is still not clear whether each motor difficulty is inherent to the disorder itself or whether they are mediated by co-occurring disorders [9, 11]. Co-occurring diagnoses are common in neurodevelopmental disorders. One often co-occurring disorder in ADHD or ASD is developmental coordination disorder (DCD), which is characterized by impaired motor coordination [12–16].

Research on motor impairments in adults with ADHD or ASD has increased during the last fifteen years [1, 17–26]. Stray et al. [19] found that muscle tone dysregulation and motor inhibition associated with ADHD may also lead indirectly to pain. Further, adults with ADHD have been shown to have impaired manual dexterity [21] and postural instability [22]. For adults with ASD, studies indicate fine and gross motor impairments [17, 23], related to, e.g. manual motor skills [24], postural stability [18, 25], and movement timing [26], as well as aberrant gait [20, 26]. In a retrospective study of adults, those with ASD or ADHD were reported to have more motor impairments in childhood compared to controls [1]. Collectively, this research has added important knowledge of motor proficiency of adults with ADHD or ASD. However, research investigating motor proficiency among those diagnosed in adulthood is scarce. Thus, more studies are needed to describe the specificity of motor impairments in adults with ADHD or ASD and to compare between subtypes of ADHD simultaneously. The present study aimed to compare (1) motor proficiency of persons diagnosed with ADHD or ASD in adulthood to normative values of motor proficiency, and (2) motor proficiency between persons with ADHD and ASD diagnosed in adulthood.

Methods

Participants

This cross-sectional study was based on the data from 153 adults with ADHD or ASD (median age 32 years, 36% women). The clinical sample consisted of persons with suspected neuropsychiatric disorders who were referred to a neurological clinic at the University Hospital of Umeå, Sweden, between 2008 and 2017. The study with a waiver of consent regarding data collected in 2008–2015 was approved by the Regional Ethical Review Board in Umeå (Reg. No. 2016-102-31M). An extension of the study, including data collected in 2016 and 2017, was approved by the Regional Ethical Review Board in Umeå (Reg. No. 2016-297-32M), and participants signed a written informed consent form one week before the test occasion by the Declaration of Helsinki. A standardized test of motor proficiency was conducted as part of a routine clinical examination done by a team consisting of a physiotherapist, an occupational therapist, a psychologist, and a physician. Exclusion criteria were moderate to profound intellectual disability (IQ < 50 according to the Wechsler Adult Intelligence Scale, WAIS-IV [27], diseases, syndromes, or conditions (e.g. cerebral palsy, severe pain) that were considered to influence the outcomes potentially. All participants were diagnosed in adulthood according to the DSM-IV criteria [28]. During the data collection, the DSM-5 [29] was published and therefore the sample was later merged into diagnostic groups according to the DSM-5. According to DSM-5 ADHD is characterized by persistent patterns of inattention and/or hyperactivity/impulsivity. In this study, the participants with ADHD were divided into two subgroups; (1) predominately inattentive presentation (ADHD-I) and (2) combined presentation (ADHD-C). The four participants with predominately hyperactive-impulsive presentation were included in the ADHD-C. For participants who had both ADHD and ASD, the individual was classified into a group based on the primary diagnosis.

Motor assessment

Participants performed the Bruininks-Oseretsky Test of Motor Proficiency, second edition (BOT-2); a standardized norm-referenced measurement to assess fine and gross motor skills [30]. The BOT-2 is considered reliable and valid to use on children and young adults up to 21 years of age [31]. The psychometric properties of using the test for older adults are however not evaluated. Since there is no gold standard assessment of adult motor proficiency, the BOT-2 was adopted here as it is a comprehensive test battery assessing many aspects of fine and gross motor skills commonly impaired in persons with ADHD or ASD. The BOT-2 consists of eight subtests: fine motor precision (FMP), fine motor integration (FMI), manual dexterity (MD), upper limb coordination (ULC), bilateral coordination (BIC), balance (BA), running speed and agility (RSA), and strength (ST). These subtests assess related aspects and are further organized into four motor composites: fine manual control (FMC), manual coordination (MC), body coordination (BC), and strength and agility (SA). The structure of BOT-2 is illustrated in the results section (Figure 1). For each subtest, the point scores are summed and converted into a scale score (mean 15 ± 5). Subtests that assess related aspects of a motor composite are combined into a standard score (mean 50 ± 10). These standard scores are summed, and this sum is converted to a Total Motor Composite (TMC, mean 50 ± 10). Scores can additionally be converted into five descriptive categories based on normative values of motor proficiency: “Well-below average (≤2SD)”, “Below average (≤1SD)”, “Average”, “Above average (≥1SD)”, and “Well-above average (≥2SD)”. Female, male, and combined standardized norms are included for analysis. Norms cover 1520 children and young adults aged 4–21 years. The standardization sample for adolescents and young adults includes 220 individuals aged 15–21 years. The norms and scoring process is described in more detail in the manual of BOT-2 [30]. Of note, the normative tables of BOT-2 for 18–21 years were applied to all participants including those older than 21 years.

Figure 1. Motor proficiency for the diagnostic groups (n = 153) in relation to standardised normative data for BOT-2 presented as proportions of test scores “within average or better” and “less than average” for each subtest, composite and total score. BOT-2 = Bruininks-Oseretsky test of motor proficiency (2nd ed.); ADHD = attention deficit hyperactivity disorder; ADHD-I = inattentive subtype; ADHD-C = combined subtype. *p < 0.05, **p < 0.01, ***p < 0.001 for each group and square brackets prior to statistical notations indicate significance for all groups combined based on Binomial test (two-tailed) for a test value of 0.8414.

Thirteen horizontal grouped bar graphs, organized in accordance with the structure of the Bruininks-Oseretsky test of motor proficiency, showing results across diagnostic groups in relation to standardised normative data for eight subscales to the left, four composites in the middle, and total composite to the right. Asterisks in each bar graph indicate significance levels for significant results, showing that motor proficiency of the three diagnostic groups combined was significantly impaired compared with normative values in all assessed motor domains except for fine motor skills.

The BOT-2 has two versions: a complete form (53 items) and a short form (14 items), of which the complete form is considered more accurate to use in children with ADHD [32]. In children without disability, the BOT-2 has demonstrated excellent internal consistency (α = 0.95–0.96), excellent interrater reliability (r = 0.98), and good test-retest reliability (r = 0.75–0.84) [30]. In children with intellectual disabilities, the BOT-2 has shown good to excellent internal consistency (α = 0.81–0.92) and test-retest reliability (ICC = 0.88–0.99), and acceptable responsiveness for all BOT-2 measures except the balance subtest [33].

Procedure

Each participant was individually assessed on one occasion by the same physiotherapist. The complete form of the BOT-2 was administered and hand-scored according to the manual. Scores from the assessments were manually entered into a data file.

Statistical analyses

Statistical analyses were performed using IBM SPSS (Statistical Package for Social Sciences, version 27). The significance level was set to α = 0.05 for all inferential tests. Missing BOT-2 item values were prevalent for 22 cases and 3% of BOT-2 item data and missing completely at random (MCAR) according to Littlés MCAR test χ² (953, n = 153) = 994.9, p = 0.17, ns. These values were imputed by a single imputation using an expectation maximum (EM) algorithm. Data distributions were assessed via visual inspection of histograms and using Shapiro-Wilks tests. Descriptive statistics are provided as means and standard deviations for normally distributed data and as medians and inter-quartile ranges (IQR) for non-normally distributed data.

The descriptive categories for the motor proficiency subtests and composites in BOT-2 were dichotomized from the original five categories from the manual, where; “Well-above average”, “Above average”, and “Average” were merged and labeled “Within average or better” and “Below average” and “Well-below average” were merged and labeled “Less than average”. The two categories are depicted as absolute and relative (% within the group) frequencies. To compare the motor proficiency of the diagnostic groups with people without disability (aim 1), the sample proportions for each subtest and composite were compared with standardized population norms using one-sample binominal tests, where 84.14% of the sample (mean motor proficiency ±1 standard deviation, or higher) was hypothesized to have a motor proficiency score in the category “Within average or better”.

To compare motor proficiency according to BOT-2 motor composite and subscale scores between the three diagnostic groups (aim 2), one-way ANOVAs and Kruskal-Wallis H-tests were used for normal and non-normally distributed data, respectively. Post-hoc comparisons using Dunn’s tests were used to explore between which groups differences were ­present. To control for type I errors in multiple family-wise testing, Bonferroni corrections were added to post-hoc comparisons.

Effect sizes for group comparisons were estimated using Eta-squared (η²). For Kruskal-Wallis tests, η²_H was calculated using the following equation: H − (k + 1)/n − k (k represents the number of groups compared), as suggested by Cohen [36]. For η² and η²_H, a coefficient of <0.01 was interpreted as negligible, 0.01–0.059 as small, 0.06–0.139 as moderate, and ≥0.14 as a large effect size by the recommendations of Cohen [34].

Results

Background characteristics for the diagnostic groups are presented in Table 1. Co-occurring neurodevelopmental disorders are specified in Supplementary Table 1.

Table 1. Characteristics of the diagnostic groups.

	ADHD-I (n = 53)	ADHD-C (n = 67)	ASD (n = 33)	Statistic (df)	P
Age, median (min–max)	28 (18–51)	34 (20–60)	32 (20–56)	9.78 (2)^a	0.008**
Women n, (% within group)	20 (38%)	22 (33%)	13 (39%)	0.53 (2)^b	0.77
DSM-IV diagnosis^c, n
ADD	43
Unspecified attention disorder	10
ADHD		63
Unspecified hyperactivity disorder		4
Asperger’s disorder			19
Autistic disorder			6
Atypical autism			8
Co-occuring diagnosis^d	6	4	14

ADD = attention deficit disorder; ADHD = attention deficit hyperactivity disorder; ADHD-I = inattentive subtype; ADHD-C = combined subtype; ASD = Autism spectrum disorder; DSM IV = Diagnostic and Statistical Manual of mental disorders (4th ed).

^aKruskal-Wallis H test; ^bPearson Chi-Square test (2 × 3 table); *p < 0.05; **p < 0.01. ^cClinical diagnoses prior to merging groups according to DSM-5. ^dCo-occuring neurodevelopmental disorders (details are presented in Supplementary Table 1).

Motor proficiency of the total sample compared with standardized norms

The proportions of persons falling within the categories “Within average or better” and “Less than average”, respectively, about standardized norms [30] are presented in Figure 1. A total of 86% (n = 132) of the sample performed less than average on one or more of the motor proficiency domains compared to norms. A higher proportion of all diagnostic groups showed poor motor proficiency compared with norms for people without disabilities. The proportions with less than average motor proficiency for the composite scores were: 41% in the TMC, 48% in the MC, 40% in the BC, 47% in the SA, and 8% in the FMC. Higher proportions of poor motor proficiency relative to norms were also found in all subscales except for FMP (8%) and FMI (8%), indicating fine manual control was within the normal range. All other subscales ranged between 38 and 58% less than average scores, most prominent in the MD subscale (Figure 1).

When comparing each diagnostic group separately to standardized norms, the ASD group did not differ significantly from the normative values for fine motor skills (FMI, FMP, FMC) (i.e., close to 84% performed “Within average or better”, p = 0.42–0.86, ns). The ADHD-I group performed better than the norm for FMP (i.e., statistically significantly more than 84% “Within average or better”, p = 0.046. The ADHD-C group performed better than the norm for FMP and FMC (both p = 0.03). For the rest of the BOT-2 subtest scores, motor composites, and total score (TMC), all groups performed significantly poorer than normative values (only 36 to 70% “Within average or better”, p < 0.001–0.006). Detailed results for all binomial tests are provided in Supplementary Table 2.

Between-group comparisons of motor proficiency

Descriptive statistics and group comparisons for the BOT-2 subtest scores, motor composites, and total scores (TMC) for the groups are displayed in Table 2. Between-group comparisons are further depicted as motor profiles for the eight subscale scores in Supplementary Figure 1, and for the standard scores of the five motor composites in Supplementary Figure 2.

Table 2. Between-group comparisons of motor proficiency according to BOT-2 subtest scores, composite scores, and total motor composite score.

	All groups	ADHD-I	ADHD-C	ASD	Group differences
Variable	(n = 153)	(n = 53)	(n = 67)	(n = 33)	Statistics^a,b	df	p	Effect size^c
Fine manual precision
Median (IQR)	19 (6)	19 (5)	19 (6)	19 (6)	0.61^a	2	0.74	<0.01
Fine manual integration
Median (IQR)	19 (5)	19 (3)	19 (5)	19 (3)	0.45^a	2	0.80	<0.01
Fine manual control
Median (IQR)	60 (9)	60 (10)	60 (11)	60 (11)	2.16^a	2	0.34	<0.01
Manual dexterity
Median (IQR)	9 (5)	9 (5)	10 (4)	9 (5)	2.82^a	2	0.24	<0.01
Upper-limb coordination
Median (IQR)	12 (7)	14 (8)	14 (7)	11 (8)	2.59^a	2	0.27	<0.01
Manual coordination
Mean (SD)	39.8 (7.3)	39.5 (8.2)	40.8 (6.8)	38.2 (6.7)	1.45^b	2, 150	0.24	0.019
Bilateral coordination
Median (IQR)	17 (8)	17 (8)	17 (7)	10 (11)	7.84^a	2	0.02*	0.039
Balance
Mean (SD)	11.0 (4.1)	11.7 (4.0)	11.2 (4.2)	9.4 (3.6)	3.44^b	2, 150	0.04*	0.044
Body coordination
Mean (SD)	42.0 (8.4)	42.8 (8.4)	43.3 (7.6)	38.0 (9.1)	4.91^b	2, 150	0.009**	0.061
Running speed and agility
Mean (SD)	12.2 (4.2)	12.0 (3.9)	13.0 (4.4)	10.9 (3.9)	3.22^b	2, 150	0.04*	0.041
Strength
Median (IQR)	10 (5)	10 (3)	12 (5)	10 (5)	8.49^a	2	0.016*	0.041
Strength and agility
Mean (SD)	41.3 (7.7)	40.4 (7.4)	43.1 (7.5)	38.9 (7.7)	4.22^b	2, 150	0.016*	0.053
Total motor composite
Mean (SD)	42.0 (6.8)	42.1 (7.0)	43.0 (6.4)	39.8 (6.8)	2.60^b	2, 150	0.08	0.034

BOT-2 = Bruininks-Oseretsky Test of Motor Proficiency (2nd ed.); ADHD = attention deficit hyperactivity disorder; ADHD-I = inattentive subtype; ADHD-C = combined subtype; ASD = Autism spectrum disorder.

^aKruskal-Wallis H test. ^bOne-way ANOVA. ^cEffect size according to η²; *p < 0.05; **p < 0.01.

Statistically significant group differences with small to moderate effect sizes were found for Body Coordination (p = 0.009, η² = 0.061) and its two related subscales Bilateral Coordination (p = 0.02, η²_H = 0.039) and Balance (p = 0.04, η² = 0.044). Also, statistically significant group differences with small effect sizes were found for Strength and Agility (p = 0.016, η² = 0.053) and its related subscales Running speed and agility (p = 0.04, η² = 0.041) and Strength (p = 0.016, η²_H = 0.041). Post-hoc tests further revealed significantly poorer Bilateral coordination, balance, and body coordination, in the ASD group compared with both the ADHD-C (H = 25.4, p = 0.006; t = 2.09, p = 0.04; and t = 3.04, p = 0.003, respectively) and ADHD-I groups (H = 20.2, p = 0.04; t = 2.64, p = 0.01; and t = 2.47, p = 0.02 respectively). For running speed and agility, group differences were only found between the ASD and ADHD (ASD < ADHD-C, t = 2.44, p = 0.02; ASD vs. ADHD-I, t = 1.34, p = 0.09, ns; ADHD-I vs. ADHD-C, t = 1.34, p = 0.18, ns). For strength and strength and agility the ADHD-C group performed better than both the ASD (H = 21.6, p = 0.02; t = 4.32, p = 0.009, respectively) and ADHD-I (H = 20.5, p = 0.01; t = 2.78, p = 0.045, respectively) groups.

After applying Bonferroni corrections to the post-hoc comparisons reported above, a few of the results changed. Group differences between ASD and ADHD-I regarding Bilateral coordination, between ASD and ADHD-C regarding balance and strength, and between ADHD-I and ADHD-C regarding Strength and agility did not remain statistically significant (p-values ranged between 0.065 and 0.14). The rest of the group differences remained statistically significant (p-values ranged between 0.009 and 0.04).

Discussion

To the best of our knowledge, this is the only study that has compared motor skill performance, measured using the BOT-2, in persons with ADHD or ASD diagnosed in adulthood. Motor proficiency of the total sample was significantly impaired compared with normative values in all assessed motor domains except for fine motor skills. For between-group comparisons, the ASD group scored significantly lower than the ADHD-I and ADHD-C groups for body coordination, with a moderate effect size. For strength and agility, the ADHD-C group scored better than the ADHD-I group as well as the ASD group, albeit with a small effect size.

The adults with ADHD or ASD displayed below-average normative performance in several motor areas. This pattern was particularly prominent for the ASD group. For fine manual control, however, performance for all groups was comparable to or better than norms. Of note, the items of fine manual control in BOT-2 are not time-based (i.e., no demands on speed and timing) and are mainly performed single-handed (i.e., no demands on bilateral coordination), which could explain the better performance of the adults in our study. An additional explanation could be that only 3% of our participants showed co-occurring DCD or intellectual disabilities. Individuals with such co-occurring conditions usually perform below average in fine manual control compared to the norms [30]. That all groups of the current study displayed problems with manual dexterity and gross motor skills compared with normative performance is in line with previous findings that motor proficiency of individuals with ADHD or ASD persists in adulthood [1, 17, 26, 35].

Adults with ASD demonstrated more difficulties in bilateral coordination and balance compared with those with ADHD. Our results are consistent with previous research concluding that impairments in motor coordination and postural control are common in ASD regardless of age [17, 19, 20, 25, 26, 36]. Cho et al. [26] recently found that adults with ASD display a specific motor signature that influences movement timing and aspects of balance. Problems with balance observed among individuals with ASD have been suggested to be due to difficulties in processing somatosensory and visual information for postural control [18]. Our results are also somewhat consistent with those of Ament et al. who found impairments in catching and static balance among children with ASD [8]. In that study, however, performance comparisons between children with ASD and ADHD were based on measures of the movement assessment battery for children [37].

The ADHD-C group performed better in strength and agility compared with both the ADHD-I and the ASD groups. In comparison to normative values of motor proficiency, all three groups performed less than average in all gross-motor composites with the lowest scores for the ASD group. Morrison et al. [35] found that as well as generally slower motor responses, adults with ASD also had decreased upper-limb strength compared to adults without disability. In that study, upper-limb strength was measured by a handgrip dynamometer. In BOT-2, upper-limb strength is measured by composite tests that require greater coordination such as push-ups. Hence, performance in Strength and Agility is not only related to strength but also to gross motor coordination skills. Therefore, lower performance in Strength and Agility in our ASD group could be explained partly by poorer coordination skills. Since low levels of physical fitness are common in adults with ASD [38], this could also be a plausible explanation for poor performance in fitness-related domains such as strength and agility in our ASD group.

The generalizability of the results from the current study is limited to persons diagnosed in adulthood with mainly mild-to-moderate motor impairments. The reason for the low number of adults with more severe motor impairments in our sample may be that such individuals are more commonly diagnosed in childhood. Another consideration is that the results of our study were based on a motor assessment developed for children and young adults aged between 4 and 21 years, with specific norms for different age groups of which the oldest is 19–21 years. Since our participants consisted of adults with a median age of 32 years (18–60 years) the results must thus be interpreted with some caution. It would of course have been preferable to use a motor assessment specifically developed and ecologically validated for adults in a wide age range. Due to the lack of a gold standard assessment of adult motor proficiency [39], the BOT-2 has been used in studies conducted on older adults with DCD [40, 41]. Finally, some of the group differences were no longer statistically significant when applying Bonferroni corrections. With modest sample sizes, corrections do however elevate the risk for type II errors and may lead to statistical differences being overlooked [42]. We, therefore, decided to favor non-corrected results, while reporting both corrected and non-corrected results for transparency.

According to our study, impaired motor proficiency is common in adults with ADHD or ASD diagnosed in adulthood. Our findings support recent research advocating motor screening, assessment, and interventions throughout the lifespan in persons with ADHD [43, 44] or ASD [45, 46]. Learning and mastering motor skills requires training and repetition from early childhood to early adulthood [47]. Given that motor skills improve with practice time and experience, we believe that people with or without disability who experience barriers early in life to participate in play and physical activities on equal terms may risk suboptimal training to develop their motor competence. Accordingly, adults with autism have reported less frequent physical activity and less positive attitudes toward physical activity [38, 48]. Identifying adults at risk of physical inactivity is thus of importance since impaired motor skills may affect both their physical and psychosocial lives. For instance, adults with ADHD and DCD are associated with increased rates of ill-health pension, welfare dependence, psychiatric disorders, and psychotropic medications [49]. Further, it is reported that older adults with ADHD are more likely to have chronic physical illness and poorer self-rated health [50] and that young adults with neurodevelopmental disorders report lower mental well-being than general populations [51]. Exercise interventions have however shown beneficial effects on physical, cognitive, behavioral, and socio-economic functions among persons with ADHD [43, 44] or ASD [52, 53]. Such positive treatment results support the importance of assessing the motor proficiency of adults with neurodevelopmental disorders and the value of improving their physical activity. Longitudinal studies are warranted to further investigate the stability of motor proficiency and the impact of aging and physical activity.

Conclusion

Our findings support clinical observations and recent research on motor impairments of adults with ADHD or ASD. Moreover, our results suggest that motor assessments should be included in clinical examinations of adults with suspected neurodevelopmental disorders. Motor assessment with BOT-2 shows potential to identify specific areas of impaired fine and gross motor skills in adults with neurodevelopmental disorders that could be of value to guide therapy and support persons with ADHD or ASD.

Acknowledgments

Andrew Strong is gratefully acknowledged for proofreading the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The study was funded by the County Council of Västerbotten.