Study reference

Study characteristics

Patient characteristics

Prognostic factor(s)

Follow-up

Outcome measures and effect size

Comments

Employment

Huang 2013a

Cross-sectional survey

Setting: Taiwan (convenience sample from CP Association members)

Source of funding: non-commercial; declare ‘no conflicts of interest’

Inclusion criteria: member of Cerebral Palsy Association; known (self-reported) diagnosis of CP; age >18 years

Exclusion criteria: none

526 individuals

279 (53%) responded to survey

Study sample:

N=279

Age

26.4 y (SD 7.7)

Sex

61% M

Types of CP

Spastic 52%

Dyskinetic 29%

Ataxia 19%

Not sure 29%

Assoc impairment

Cogn disab 28%

Speech diff 18%

Education

College or above 19%

Senior HS 52%

Junior HS 19%

Elem school 9%

Anonymous mailed survey, followup telephone call to increase response rate

Outcome = employment = full-time or part-time in integrated employment or sheltered employment settings, excluding unpaid work

Analysis: multivar logistic regression of employability; 2 sets of covariates were entered

in sequential steps: (a) demographic and educational variables (gender, age, types of CP, educational attainment, types of school attended); and (b)

mobility in community and independence in ADL

Endpoint of follow-up: cross-sectional

For how many participants were no complete outcome data available?

N (%): 47% (non-responders)

Reasons for incomplete outcome data described? No (no followup on non-responders!)

Predictors of employability; OR [95% CI]

Employability: employed = full-time or part-time in integrated employment or sheltered employment settings, excluding unpaid work

Educational attainment on 4-point scale (elementary to college or above); type of school on 4-point scale (special school to regular); mobility in community (typical performance)  in 4 levels (by others, wheelchair, walk with assist device, walk without assist device); level of independence in 6 levels (complete dependence to total independence)

Full model (OR [95% CI])

Gender, Man 1.21 [0.60;2.42] NS

Age 1.05 [1.01;1.10]

Types of CP (spastic = ref)

Dyskinetic 1.14 [0.45;2.86] NS

Ataxia 3.44 [1.29;9.13]

Not sure 1.68 [0.66–4.31] NS

Educ attainment 1.86 [1.09;3.18]

Types of school 1.00 [0.75;1.33] NS

Mob in community 1.48 [1.04;2.10]

Indep in ADL 1.60 [1.23;2.09]

Authors conclude that employment rate for adults with CP in Taiwan is low (22.9%); age, CP diagnosis, educational attainment, and functional performance are important determinants related to employment outcomes

‘Results underscore importance of functional performance on employment outcomes. May also (in part) be due to employers’ misperception.’

‘Rehabilitation should focus on self-care skills, community mobility, and independence training to improve the ability of individuals with CP to demonstrate employability to employers. The removal of environmental barriers and the provision of assistive technologies to increase independence should be emphasized.’

Note: in contrast with several other studies, individuals with cognitive problems, walking limitations etc. are not excluded, i.e. study sample represents a full range of abilities

Note: relatively low employment rate (22.9%), lower than Europe / US; study in Taiwan; convenience instead of random sample, i.e. only responders were analyzed (who are generally more active in seeking resources to deal with their problems); study based on self-report (i.e. without physical examination); study population with relatively high mean age (26 years)

Note: the increased likelihood of employment in case of ataxia may be due to better gross motor function in this type of CP

Huang 2013b

Retrospective cohort (registry analysis)

Setting: USA (state vocational rehabilitation services)

Source of funding: non-commercial

Inclusion criteria: clients with CP receiving state vocational rehabilitation services with an individualized plan for employment (RSA-911 database); cases closed in 2009 (as succesful employment, or as unemployed)

Exclusion criteria: none

N=3162

Age

16-54 years

Sex

58% M

Race

Europ-Am 75%

Educ level

Special edu 14%

<high school 29%

High school 31%

Post-sec 9%

Bachel or high 9%

Intell disabil 9%

Epilepsy 3%

Med insur 54%

Cash benefits 60%

RSA-911 database: dataset for cases closed in 2009

Outcome = Competitive employment = working full time or part time in integrated competitive setting, in self-employment or in a state-managed business enterprise programme with an income compensated at or above minimum wage (unsuccessful outcome = unable to find employment after receiving vocational rehabilitation service interventions).

Analysis: multivar logistic regression of employability; demographic variables, disincentive variables, rehabilitation service variables; 2 analyses: (1) demographic + disincentive variables; (2) demographic + vocational rehabilitation variables

Endpoint of follow-up: 41.2 months (mean duration between eligibility and case closure)

For how many participants were no complete outcome data available?

N (%): none

Reasons for incomplete outcome data described? Not relevant

Descriptioin of vocational rehabilitation services:

On-the-job training Training in specific job skills by a prospective employer

Job placement assistance A referral to a specific job resulting in an interview

On-the-job support Support services provided to an individual who has been placed in employment in order to stabilize the placement and enhance job retention

Maintenance Monetary support provided for those expenses such as food, shelter, and clothing

Rehabilitation technology The systematic application of technologies, engineering methodologies, or scientific principles to meet the

needs of, and address the barriers confronted by, individuals with disabilities

Predictors of employability; OR [95% CI]

Employability: employed = competitive employment, excluding unpaid work

49.6% closed in 2009 as succesful employment; 50.4% as unemployed [note: employment rate of people without disabilities is 69%]

Full model (OR [95% CI])

Male sex

1.27 [1.09–1.48]

Age (3 categories; 26–54y = ref)

16–20y 0.68 [0.56–0.81]

21–25y 0.86 [0.69–1.08] NS

Race (European-Am = ref)

1.14 [0.92–1.42] NS

Edu level (<bachelor degree =ref)

2.74 [2.02–3.71]

Med insurance (‘No’ = ref)

0.89 [0.77–1.03] NS

Cash benefits (‘No’ = ref)

0.55 [0.45–0.66]

Diagn and treatm (‘No’ = ref)

1.18 [1.00–1.39] NS

On-the-job training (‘No’ = ref)

1.53 [1.04–2.25]

Job placem assist (‘No’ = ref)

2.80 [2.39–3.28]

On-the-job support (‘No’ = ref)

2.33 [1.93–2.80]

Maintenance (‘No’ = ref)

1.51 [1.21–1.87]

Rehab technol (‘No’ = ref)

1.80 [1.48–2.18]

Authors conclude that males aged between 26 and 54 years old with higher education attainment were more likely to be employed; individuals receiving disability benefits were less likely to be employed. After controlling for demographic and work disincentive variables, 5 vocational rehabilitation services significantly predicted employment outcomes (p<0.05): on-the-job training; job placement assistance; on-the-job support; maintenance services; and (5) rehabilitation technology.

Note: US study on individuals receiving state vocational rehabilitation services; may not be completely representative of situation in NL (e.g. not receiving disability benefits associates with having employment)

Note: negative association with disability related benefits may be due to relation with functional severity. Obtaining these benefits may be a proxy for severity which could more directly be linked to unemployment.

Note: prolonged needs for education and vocational training may reduce employment rate for young adults

Note: RSA-911 database may suffer from subjective recall bias; database does not contain specific disability and service information related to CP, such as severity, types of CP, physical function etc ..

Note: authors state that their study has a cross-sectional design (..)

Bjornson 2011

Prospective cohort

Setting: USA (sample of youth from national longitudinal transition study))

Source of funding: non-commercial (US department of Education)

Inclusion criteria: students 13-16 years, in seventh grade or higher on Dec 1, 2000; orthopedic impairment

National Longitudinal Transition Study 2 (NLTS2) = nationally representative sample of students(received special education services at entry in the study), info collected prospectively in 5 waves as they transitioned from adolescence in secondary school to early adulthood

Current analysis = 2nd analysis NLST2 using info wave 1 (2000-2001) and wave 4 (2006-2007): participants with ‘orthopedic impairment’ who completed wave-1

N=1510 (completed wave-1)

Analytic sample:

N=990 (completed wave 1 + 4)

Age

15.1 (wave-1)

21.1 (wave-4)

Sex

57% M

Impairment

53%  CP

9% SB

Trauma 4%

Multiple 2%

Other 32%

Race

White 63%

Afr-Am 16%

Hisp/Latin 17%

Educ head household

< high school 13%

High school 30%

Some college 27%

Bachel or high 30%

Health insurance

Any 57%

Data were collected from a variety of sources, including parents and guardians, teachers, principals, school records (wave 1+5), and the students (wave 4)  themselves.

Outcome = employment: worked for pay in last 2 years or has a paying job’(additional outcome = ‘any type of postsecondary education’) [wave-4]

Main predictor = OT/PT: ‘received PT or OT in last 12 months, at age 13-16 years’ [wave-1]; School-based and

non–school-based OT and/or PT were provided to 34% of

participants at wave 1

Analysis: multivar logistic regression of employability (and post-sec education) adjusted for gender, ethnicity, age, any social interactions (dichotomous), parental education, expressive language (proxy for cognition), wheelchair mobility, any health insurance, received a high school diploma/certificate, overall social skills, and percent students attending schools where more than 50% qualified for reduced or free lunch (as a proxy for community SES)

Expressive language =  average of

speaking and communicating variable on 4-point Likert

scale (1 = speaks/communicates as clearly as other chil-

dren/no trouble communicating to 4 = does not speak or

communicate at all)

Note: NLST2 uses complex survey sampling method >> sampling weights were used in analyses to correct for sampling bias

Endpoint of follow-up: 6 years (mean duration between wave 1 and 4)

For how many participants were no complete outcome data available?

N (%): 520 (34%)

Reasons for incomplete outcome data described? Not described

Predictors of employability; OR [95% CI]

Employability: employed = worked for pay in last 2 years or has a paying job’ [wave-4]

Full model (OR [95% CI])

43% current or prior employment in last 2 years; 20% current employment

any OT or PT

0.44 (0.87-5.82) NS

Female gender

1.8 (0.83-3.82) NS

Nonwhite ethnicity

0.51 (0.35-0.75)

Age (per year)

0.79 (0.58-1.08) NS

any social interactions

12.2 (2.22-66.87)

Higher parental education

1.3 (0.89-1.86) NS

increased expressive language score (higher score = worse expression!)

0.50 (0.35-0.72)

Wheelchair use

0.06 (0.02-0.14)

health insurance

0.50 (0.49-1.33) NS

Child HS diploma or GED

4.2 (1.77-9.99)

increased overall social skills

1.1 (0.97-1.12) NS

Attends a school with higher

percent of students qualifying

for reduced price or free lunch.

0.85 (0.62-1.2) NS

Subgroup analysis for participants aged 21 or older [who no longer qualify for special education services]: 

statistically significant predictors only

Nonwhite ethnicity

0.08 (0.04-0.16)

Increased expresslanguage

5.0 (2.93-8.52) [see Note ..!]

wheelchair use

51.4 (17.7-149.7)

Child HS diploma or GED

0.14 (0.04-0.48)

PS. if postsecondary education is used as outcome (dependent variable), OT/PT associate with increase in report of postsecondary education: OR= 3.2 [1.13; 9.15]

Authors conclude that receiving physical and/or occupational therapy

at ages 13 to 16 years was significantly associated with higher levels of enrollment in postsecondary education at ages 19 to 21 years; social interactions and expressive language skills but not therapy services were associated with paid employment (Longer follow-up is needed to effectively examine paid employment).

‘individual’s experience with social interactions, level of social skills, and/or acquisition of a high school

diploma (GED) enhanced the likelihood of paid employment’

Note: study on youth with ‘orthopedic impairment’; only 53% CP

Note: secondary analysis (post hoc); intensity and/or specific type PT/OT services provided to participants is not available, and there is no info on PT/OT at younger age (<13 years); US study may not be completely representative of the situation in NL

Note: in scoring of expressive language, a higher score = worse expression (!); there seems to be a discrepancy between primary analysis and subgroup analysis (OR=0.5 vs OR=5.0) >> see main text which is not in line with OR=5.0 for the subgroup aged 21 up (..) >> OR should be 1/5=0.2 for this subgroup ??!

Note: idem for wheel chair use (OR=0.06 vs OR=51.4]; for the older subgroup, the text states ‘wheelchair use may be a marker of increased independence in life, or users are more impaired and therefor more likely to work in supportive environment [but how to explain the discrepancy with the primary analysis with OR=0.06 ..?]

Nieuwenhuijsen 2009

Cross-sectional

Setting: rehabilitation centers Netherlands

Source of funding: non-commercial

Inclusion criteria: diagnosis of CP, age 16-20 years, normal intelligence (IQ >70)

Exclusion criteria: severe learning disabilities (IQ<70); additional diagnosis with lasting effects on motor functioning; insufficient knowledge of Dutch language

CP Transition Study South West NL at 2-year follow-up (T2)

recruited from 8 participating rehabilitation centers and rehabilitation departments in NL; 437 participants: 35% excluded because of severe learning disabilities, 18% due to incorrect address >> 184 were send information letter: 103 participated (response rate = 56%)

87/103 participated in 2-year follow-up

N=87

Age

19.9 ± 1.4

Sex

59% M

CP specs

49%  hemi

31% diplegic

20 quadri

94% spastic CP

GMFCS I/II/III/IV/V

72/10/6/9/1%

MACS I/II/III/IV/V

77/12/6/1/1%

Educational level

27% low

40% middle

33% high

Student/employment

61% student

20% paid job

19% other

Housing status

70% with parents

22% alone or w others

8% in institute

Trained interviewers: COPM; linked to ICF; personal and CP-related characteristics, incl sex, age, limb distribution of the paresis, level of gross motor functioning (GMFCS)  and manual

ability (MACS)

Outcome = experienced problems with paid or unpaid work

Analysis: simple associations between experienced problems and personal and CP-

related characteristics: Phi coefficients (Phi) for 2 dichotomous variables, Pearson correlation coefficients (Rp) for associations between dichotomous and continuous variables, and Spearman correlation coefficients (r) for associations between dichotomous and ordinal variables.

At significance level of 0.01

Endpoint of follow-up: 2 years

For how many participants were no complete outcome data available?

N (%): 16 (16%)

Reasons for incomplete outcome data described? Yes; loss of interest (9), competing

time demands (5), or relocation to another country (2).

Drop-outs had lower educational level (39% vs 27% with low

level of education)

Predictors of experienced problems with acquiring, keeping, terminating a job, or seeking employment; Spearman correlation coefficient

Employment: experienced problems with acquiring, keeping, terminating a job, or seeking employment (paid or unpaid)

Simple associations (Spearman)

11/87 (13%) experienced  problems with employment

Sex

0.05 NS

Age

-0.02 NS

GMFCS Level

0.10 NS

MACS Level

0.07 NS

Level Education

-0.30 p=0.005  [stat. sign at p<0.01]

Level of education was classified as low (prevocational practical education or less), medium (prevocational theoretical education and upper secondary vocational education), or high (secondary nonvocational education, higher education, and university)

Authors conclude that problems in functional mobility, paid or unpaid work, and socialization were considered as most important (represented by the highest mean importance score).

‘during the transition into adulthood, young adults with CP may experience problems regarding domestic life and work, which they consider important’

Note: outcome measure = self-reported (experienced) problems with employment

Note:cross-sectional desing and simple statistics i.e. without adjusting for (potential) confounders

Note: relatively small number of

participants reported  work problems; this may reflect young age

of the participants and that they may still be in school rather than working; problems arose mainly in participants with lower levels of education, who are the ones who had finished their education and are now working or looking for work.

Note: power of the study at alpha of 0.01 was 0.90 to detect associations of 0.40 or greater (..)

Magill-Evans 2008

Cross-sectional

Setting: Canadian province (urban and rural area) ; non-random sample

Source of funding: non-commercial

Inclusion criteria: diagnosis of CP or SP; willingness to participate

Non-random sample: recruited through service providers, previous studies,

newspaper advertisements, and referrals from other participants

N=76

Age

25.0 ± 3.1 (20-30 y)

Sex

46% M

GMFCS (mean)

2.7 ± 1.3

IQ

85.9 ± 17.1

Educational (mean)

2.7 ±  1.6

Transportation independence

1.7 ±  0.8

Transition services

1.0 ± 1.0

mailed questionnaires to complete prior to a semi-structured interview

Outcome = level of employment: 7-point ordinal scale with competitive full-time (code 6),competitive part-time, noncompetitive paid, noncompetitive unpaid, not employed currently but in the past, and never been employed (coded 0)

Analysis: Multiple regression of individual, family, and community variables to identify those that accounted for a statistically significant amount of variance in employment.

Model 1: age, gender, GMFCS, IQ,

education, task coping, and depression; significant variables carried forward into Model 2: added the three family variables; significant variables carried forward into Model 3: added the two communitylevel variables (full model)

At significance level p<0.05

Endpoint of follow-up: cross-sectional

For how many participants were no complete outcome data available?

N (%): 1-6 ‘missing cases’ (<10%) depending on variable

Reasons for incomplete outcome data described? No, not specified

Predictors of employment status; regression coefficients (Beta)

Employment status: 7-point ordinal scale with competitive full-time (code 6),competitive part-time, noncompetitive paid, noncompetitive unpaid, not employed currently but in the past, and never been employed (coded 0)

Full multivar model (model 3)

Employment status, n (%)

6: 17/87 (19.5%)  [comp full time]

5: 18/87 (20.7%)

4: 3/87 (3.4%)

3: 3/87 (3.4%)

2: 19/87 (21.8%)

1: 16/87 (18.4%) [never employed]

Gender(male=ref)

−0.33 (p<0.05)

GMFCS

−0.18 NS

IQ

0.27 (p<0.05)

Education

–  NS

Task coping

–  NS

Depression

–  NS

Family functioning

–  NS

Father fosters autonomy

–  NS

Mother fosters autonomy

–  NS

Transportation Independence

0.23 (p<0.05)

Transition services

0.02 NS

Model-3 explains 42% of the variance in employment

Authors conclude that gender (females were less likely to be employed), IQ (lower IQ associated with unemployment), and transportation dependence

accounted for 42% of the variance in employment

‘Transportation is a significant barrier to employment and innovative solutions are needed. Issues related to gender

need to be considered when addressing employment inequities for persons with primarily motor disabilities’

Note: education not a sign predictor of employment (unlike previous studies) >> may be explained by inclusion of participants with low IQ and/or relatively young age of participants

Note: non-random sample including volunteers (via advertisements) may be non-representative

Note: small sample size relative to the number of predictors in the regression models

Donkervoort 2007

Cross-sectional

Setting: rehabilitation centers Netherlands

Source of funding: non-commercial

Inclusion criteria:

Exclusion criteria:

See Nieuwenhuijsen 2009

CP Transition Study South West NL at baseline (T1)

N=103

Age

17.9 ± 1.4

Sex

60% M

CP specs

49%  hemi

25% diplegic

26 quadri

86% spastic CP

GMFCS I/II/III/IV/V

76/7/5/12/1%

Educational level

39% low

34% middle

25% high

Student/employment

85% student

10% paid job

Housing status

83% with parents

14% independent

3% in institute

Socio-econ stat (SES)

11% low

48% medium

33% high

3 standardized measurment instruments on ADL and particpation: LIFE-H, FIM/FAM, and VABS

Outcomes = (1) problems with employment: LIFE-H subdomain Employment; (2) relationships: LIFE-H and VABS subdomain interpersonal relationships

Analysis: multiple regression adjusting for gender, age, GMFCS level, CP limb distribution, level of education and SES; stepwise (entry 0.05, exit 0.10); p<0.05

Endpoint of follow-up: cross-sectional

For how many participants were no complete outcome data available?

N (%): none (but note that original response rate was 56%: see Nieuwenhuijsen 2009; no sign diff between participants and non-responders)

Reasons for incomplete outcome data described? Not relevant

Predictors of problems with employment; regression coefficients (Beta)

(1) Employment: LIFE-H subdomain Employment

(2) Relationships: LIFE-H and VABS subdomain interpersonal relationships

Multivariate model

(1) Employment

GMFCS Level

0.47

[explains 22% of variance]

LIFE-H subdomain Employment only applicable to 40 participants: 30% reported problems with respect to employment (LIFE-H)

(2) Relationships:

LIFE-H (interpersonal relationships)

GMFCS Level

0.38

[explains 14% of variance]

VABS (interpersonal relationships)

GMFCS Level

0.33

Educ 0.25

[explain 21% of variance]

Authors conclude that 20–30% encountered restrictions in ADL (mobility, self-care, nutrition) and social participation (taking responsibility, community living, leisure activities and employment); GMFCS level, level of education, and age proved to be important determinants of functioning in ADL and social participation, explaining 70% and 66% of the variance in outcome respectively; GMFCS level associates with (problems with) employment

Note: outcome measure = self-reported (experienced) problems with employment

Note:  10% of participants have a

paid job and 14% live on their own. These percentages are low compared with the reference value of the Dutch population of this age with an employment rate of 32% and 25% living independently

Intimate relationships

Wiegerink 2012

Cross-sectional analysis

Setting: rehabilitation centers Netherlands

Source of funding: non-commercial

Inclusion criteria:

Exclusion criteria:

See Nieuwenhuijsen 2009

CP Transition Study South West NL at 4 year follow-up (T3)

28% of the 103 initial participants dropped out

N=74

Age

20-25 years

Sex

62% M

GMFCS I/II/III/IV/V

76/8/5/9/1%

MACS I/II/III/IV/V

85/11/3/0/1

Educational level

20% low

40% middle

40% high

CP specs

49% unilat

51% bilat

Speech

91% audible

8% audible / disturbed

1% disturb / not audib

Trained physiotherapists; GMFCS, MACS, paresis distribution, audibility, Questionnaire of the Dutch National Study on Sexuality, General Self-Efficacy Scale-12 (GSES-12), Dutch Personality Questionnaire (NPV), Physical Disability Sexual and Body Esteem Scale (PDSBES), environmental factors, Mobility outdoors measured by the Impact on Participation and Autonomy (IPA), parenting style assessed using short form of the EMBU, social support assessed with the Social Support List 12-Interaction (SSL-12)

study focused on “having a current romantic relationship” as an outcome parameter for romantic relationships, and “experience with cuddling nude”

and “experience with intercourse” as outcome parameters for sexual activity

Data of the Dutch National study used as a reference sample for the sexual milestones

Outcomes = “a current romantic relationship”, “experience with cuddling nude” and “experience with intercourse”

Analysis:

multivariate logistic regression analyses (Forward Likelihood Ratio); dichotomized results based on median of study sample (low=0; high=1 ; stepwise: basic model including associated demographic and physical characteristics; personal factors or environmental factors added

Endpoint of follow-up: cross-sectional analysis (at 4 years followup)

For how many participants were no complete outcome data available?

N (%): 29/103 (28%) initial participants dropped out

Reasons for incomplete outcome data described? Yes; loss of interest (n = 13), competing time demands (n = 5), incorrect

addresses (n = 5), reason unknown (n = 3) or relocation to

another country (n = 1)

There were fewer participants with a low education level at T3 (39% versus 30%, p = 0.02).

Predictors of romantic relationships and sexual activity; OR [95% CI]

(1) Current romantic relationship

(2) Cuddling nude

(3) Intercourse

Multivariate models

(1) current romantic relationship

Statist sign predictors only

Gender (M=0) 3.1 [1.1-9.0]

Self-efficacy

Effort 5.35 [1.5–19.7 ] R2= 0.20

Personality

Self-esteem 5.2 [1.5–18.6] R2= 0.21

Sex-esteem 5.7 [1.5–21.6] R2= 0.22

Sec (special) school

7.1 [0.9–58.3] R2= 0.11

Parenting style

5.5 [1.1–26.3] R2= 0.12

(2) Cuddling nude

Statist sign predictors only

Age 1.5 [1.0–2.3]

GMFCS 2.4 [1.2–4.8]

Personality

Soc adequacy 9.7 [2.6–37.1] R2=0.35

Self-esteem 3.8 [1.2–11.9] R2= 0.21

Sex-esteem 8.8 [2.4–32.3] R2= 0.31

Mobility 7.2 [1.3–38.9] R2= 0.12

Sec (special) school

3.8 [1.1–12.5] R2=0.10

(2) Intercourse

Statist sign predictors only

Age 1.6 [1.1–2.3]

GMFCS 2.4 [1.2–5.0]

Self-efficacy

Initiative 5.7 [1.8–17.9] R2= 0.25

Personality

Self-esteem 3.8 [1.3–10.9] R2= 0.21

Sex-esteem 5.3 [1.7–16.6] R2= 0.32

Sec (special) school

3.5 [1.0–11.6] R2= 0.09

Authors conclude that more females than males with CP were in a current

romantic relationship. Self-esteem, sexual esteem and feelings of competence regarding self-efficacy contributed positively to having current romantic relationships. A negative parenting style contributed negatively

Age and gross motor functioning explained 20% of the variance in experience with intercourse. In addition, sexual esteem and taking initiative contributed significantly to intercourse experience

For young adults with CP personal factors (20–35% explained variances) seem to contribute more than environmental factors (9–12% explained variances) to current romantic relationships and sexual experiences.

‘We advice parents and professionals to focus on self-efficacy, self-esteem and sexual self-esteem in development of young adults with CP’

Note: relatively small sample size; patients with IQ<70 excluded; restricted to age 20-25 years; broad confidence intervals; multiple testing

Note: most participants had only slight limitations in gross motor functioning >> study gives limited understanding of consequences of severe limitations in gross motor functioning for sexual activities; results can also not be generalized

to older people with CP

Wiegerink 2010a

Cross-sectional analysis

Setting: rehabilitation centers Netherlands

Source of funding: non-commercial

Inclusion criteria:

Exclusion criteria:

See Nieuwenhuijsen 2009

CP Transition Study South West NL at 2 year follow-up (T2)

16% of the 103 initial participants dropped out

N=87

Age

18-22 years

Sex

59% M

GMFCS I/II/III/IV/V

78/9/6/6/1%

MACS I/II/III/IV/V

84/9/5/1/1

Educational level

27% low

41% middle

32% high

CP specs

49% unilat

51% bilat

GMFCS, MACS, Vineland Adaptive

Behaviour Scales, Questionnaire of the Dutch National Study on Sexuality

Data of the Dutch National study used as a reference sample for the sexual milestones

Outcomes = “’going out’, ‘dating experience’, ‘romantic relationship experience’,

and ‘sexual activity’ (dichotomized)

Analysis:

Stepwise multivariate logistic regression analyses

Endpoint of follow-up: cross-sectional analysis (at 2 year followup)

For how many participants were no complete outcome data available?

N (%): 16/103 (16%) initial participants dropped out

Reasons for incomplete outcome data described? No

‘No significant non-response bias or lost to followuup bias was observed’

Predictors of romantic relationships and sexual activity; OR [95% CI]

(1) Dating

(2) Romantic relationship

(3) Sexual activity

Multivariate models

(1) Dating

Statist sign predictors only

Age 1.6 [1.1-2.4]

Peer group activities

Nr friends (<6=ref)

4.1 [1.4-11.7] R2=0.2

Going out (<2=ref)

73.6 p=0.01 R2=0.25

(2) Romantic relationship

Statist sign predictors only

Laterality (uni=ref)

0.29 p=0.03

Peer group activities

Mixed peers (same sex=control)

3.17 p=0.05 R2=0.16

Sporting (yes=ref)

0.22 p=0.01 R2=0.21

Going out (<3=ref)

4.2 [1.3-12.9] R2=0.12

Preceding stages

Dating (no=ref)

22.8 [5.9-88.5] R2=0.47

(3) Sexual activity

Statist sign predictors only

Age 2.1 [1.4-3.2]

Peer group activities

Nr friends (<6=ref)

1.7 p=0.03 R2=0.28

Going out (<3=ref)

4.6 [1.6-13.2] R2=0.33

Preceding stages

Dating (no=ref)

107.8 [14.7-788.5] R2=0.68

Romantic relationships (no=ref)

118.2 [11.2-1251.9] p<0.001

Authors conclude that young adults with CP had less experience in romantic and sexual relationships than an age-appropriate Dutch reference population. Peer group activities and dating favoured development of romantic relationships and sexual activity. Older age was associated with greater sexual activity. Motor functioning, education level, and gender did not correlate with romantic relationships or sexual activity

‘Being involved in peer group activities and creating a context to arrange dates seems relevant for young adults with CP to develop romantic relationships and sexual activity’

Note: relatively small sample size; patients with IQ<70 excluded; restricted to age 18-22 years; very broad confidence intervals

Note: most participants had only slight limitations in gross motor functioning >> study gives limited understanding of consequences of severe limitations in gross motor functioning for sexual activities; results can also not be generalized

to older people with CP

Wiegerink 2010b

Prospective cohort

Setting: rehabilitation centers Netherlands

Source of funding: non-commercial

Inclusion criteria:

Exclusion criteria:

See Nieuwenhuijsen 2009, and Wiegerink 2010a, 2012

CP Transition Study South West NL at baseline (T1), and at 2 year (T2) and 4 year follow-up (T3)

GMFCS, Vineland Adaptive

Behaviour Scales, Questionnaire of the Dutch National Study on Sexuality

Data of the Dutch National study used as a reference sample for the sexual milestones (: 912 men and 901 women aged 20-24 years in 2005); 4 sexual milestones: French kissing, feeling and caressing under the clothes, cuddling nude, intercourse

Outcomes = dating experience,

current romantic relationship, and intercourse experience, (dichotomized)

‘Dating experience can be a

preceding stage in the development of a romantic relationship and/or sexual activity. A current romantic relationship represents the present situation, and intercourse is the last sexual

milestone’

Analysis:

repeated-measures analyses (T1,T2,T3) for dating experience, current romantic relationship, and intercourse experience; time as an independent factor along with 1 additional factor (gender, educational level, or GMFCS level at t1); evaluate each factor’s interaction effect with time; If no interaction, then main effects of the factor and time were estimated; level of significance at p 0.05; missing values were mputed

Endpoint of follow-up: 4 year followup (repeated measures analysis; 0, 2 and 4 year followup)

For how many participants were no complete outcome data available?

N (%): at T2, 16/103 (16%) initial participants dropped out; at T3 27/103 (26%) initial participants dropped out

Reasons for incomplete outcome data described? Yes

At t3: loss of interest (13), competing time demands (5), incorrect addresses (5), reason unknown (3), or relocation to another country (1).

‘No significant nonresponse or lost-to-follow-up biases were observed in terms of age, gender, gross motor functioning, and limb distribution of paresis’

Compared with the t1 (39%) group, there were fewer participants with a low education level at t2 (33%; p=0.03) and t3 (30%; p=0.02)

Predictors of romantic relationships and sexual activity; p-values (repeated measures analysis)

(1) Dating

(2) Romantic relationship (current)

(3) Sexual activity (intercourse)

Bivariate repeated measures analysis

(1) Dating

Interaction effects

Gender*time p=0.06 NS

Edu level*time p=0.02

GMFCS*time p=0.2 NS

Main effects

Gender p=0.48 NS

Time p<0.001

Edu level NA (not applicable)

Time NA

GMFCS p=0.13 NS

Time p<0.001

interaction edu level and time: young adults with lower edu levels began dating later than those with higher education

(2) Romantic relationship

Interaction effects

Gender*time p=0.46 NS

Edu level*time p=0.34 NS

GMFCS*time p=0.34 NS

Main effects

Gender p<0.001

Time p=0.19 NS

Edu level p=0.18 NS

Time p=0.17 NS

GMFCS p=0.90 NS

Time p=0.21

main effect gender: significantly more women in current relationships than men

(3) Sexual activity (intercourse)

Interaction effects

Gender*time p=0.36 NS

Edu level*time p=0.58 NS

GMFCS*time p=0.78 NS

Main effects

Gender p=0.44

Time p<0.001

Edu level p=0.07 NS

Time p<0.001

GMFCS p=0.02

Time p<0.001

GMFCS level (I–II vs III–V) was

associated significantly with intercourse experience (but not

with dating)

Authors conclude that young adults with a lower education level

began dating later than those with higher levels. Significantly

more women were in current romantic relationships than men.

Level of gross motor function was associated significantly with intercourse experience [but not with dating and romantic relationships]. Compared with an age-appropriate Dutch reference population, young adults with CP participated at a lower level in romantic relationships and sexual activities, but had equal sexual interest at the final assessment.

‘similar sexual interests and increasing experiences with romantic

relationships and sexual activities during transition to young adulthood. However, percentage

of young adults with CP in current romantic relationships was low, especially for men’

‘health care professionals must include discussions of sexuality; the Rotterdam Transition Profile

can be used, and HEADSS (home, education, activities, drugs, sex, and suicide); discuss challenges of romantic relationships and sexual activity’

Note: at T3 as compared to Dutch reference group (Table 2) >>

Sexual interest

Fantasiz on girls: 66 vs 54% (p=0.06)

Fantasiz on boys: 33 vs 41% (p=0.1)

Masturbating: 80 VS 85% (p=0.19)

Romantic relationships

Dating: 76% vs not determined

Exper with romantic relationships: 79 vs 91% (p<0.001)

Current romantic relationship:

28 vs 63% (p<0.001)

Sexual activity (4 milestones):

French kissing: 83 vs 91% (p<0.001)

Caressing: 76 vs 90% (p<0.001)

Cuddling nude: 70 vs 83% (p<0.001)

Intercourse: 60 vs 83% (p<0.001)

In comparison with the reference population, young adults

with CP showed significantly delayed development for all

relational aspects. Many young people with CP reached sexual milestones significantly later than the Dutch reference group [but had equal sexual interest at T3]

Note: post hoc analysis shows less dating experience in subgroup of participants with IQs of 70-85

Auteur en jaartal

Redenen van exclusie

Tsybina 2012

Geen (SR van) origineel onderzoek

Watson 2011

Verkeerde PICO; geen antwoord op de uitgangsvraag

Darrah 2010

Geen vergelijkend onderzoek; geen antwoord op de uitgangsvraag

Horsman 2010a

Narrative review

Horsman 2010b

Geen vergelijkend onderzoek

Trovato 2010

Verkeerde PICO; betreft ‘equipment transition’

Donkervoort 2009

Verkeerde PICO; geen antwoord op uitgangsvraag

Gorter 2009

Narrative review

Haak 2009

Narrative review

Roebroeck 2009

Narrative review

Binks 2007

Geen antwoord op de uitgangsvraag door gebrek aan prospectieve en kwantitatieve data

Livingston 2011

Verkeerde PICO; geen antwoord op de uitgangsvraag

Palisano 2011

Verkeerde PICO; geen antwoord op uitgangsvraag

Kang 2010

Verkeerde PICO; geen antwoord op uitgangsvraag

Mandelco 2010

Narrative review

Koritsas 2009

Verkeerde PICO

Van Meeteren 2008

Verkeerde PICO (uitkomstmaat)

Wiegerink 2008

Verkeerde PICO; geen antwoord op uitgangsvraag

Hilberink 2007

Geen vergelijkende studie; verkeerde PICO

Krakovsky 2007

Geen vergelijkende studie; verkeerde PICO

Michelsen 2006

Verkeerde PICO

Shields 2006

Verkeerde PICO

Wiegerink 2006

SR van mindere kwaliteit; beslaat literatuur tot 2003 en is daarmee minder relevant

Young 2006

Verkeerde PICO

Gannotti 2013

Verkeerde PICO

Longo 2013

Verkeerde PICO

Kent 2013

Geen systematisch review

Reddihough 2013

Verkeerde PICO; geen antwoord op uitgangsvraag

Wiegerink 2006

Yes

Yes

Unclear*1

Unclear*1

Unclear*1

No*1

No*1

No*1

No*1

Study reference

(first author, year of publication)

Was there a representative and well-defined sample of patients at a similar point in the course of the disease?

(yes/ no/ unclear)

Was follow-up sufficiently long and complete?

(yes/ no/ unclear)

Was the outcome of interest defined and adequately measured?

(yes/ no/ unclear)

Was the prognostic factor of interest defined and adequately measured?

(yes/ no/ unclear)

Was loss to follow-up / incomplete outcome data  described and acceptable?

(yes/ no/ unclear)

Was there statistical adjustment for all  important prognostic factors?

(yes/ no/ unclear)

Level of evidence

Huang 2013a

Unclear*1

Cross-sectional

Yes

Unclear*1

Unclear*1

Yes

B

Huang 2013b

Unclear*2

Unclear*2

Yes

Yes

Unclear*1

Yes

B

Bjornson 2011

Yes

Unclear*3

Unclear*3

Unclear*3

No*3

Yes*3

B

Nieuwenhuijsen 2009

Yes*4

Cross-sectional

Yes*4

Yes

Yes

No*4

B

Magill-Evans 2008

Unclear*5

Cross-sectional

Yes

Yes

Yes

Yes

B

Donkervoort 2007

Yes*6

Cross-sectional

Yes*6

Yes

Yes *6

Yes

B

Wiegerink 2012

Yes*7

Cross-sectional

Yes

Yes

Yes

Yes*7

B

Wiegerink 2010a

Yes*8

Cross-sectional

Yes

Yes

Yes

Yes*8

B

Wiegerink 2010b

Yes*9

Yes

Yes

Yes

Yes

No*9

B