Pediatric Chronic Illness and Developmental Milestones
Pediatric Chronic Illness and Developmental Milestones
The present study reports results from the first broad meta-analysis on the achievement of developmental milestones of emerging/young adults with and without chronic illness and/or physical/sensory disability. Rates of success in individuals with chronic illness and disability were between 38 (parenthood) and 22% (leaving home) lower compared with their healthy and nondisabled peers. Reductions of success rates were more pronounced if individuals had a neurological disease or sensory impairment and if the disease or disability was highly visible to others. In addition, effects of chronic illness and disability on economic independence became smaller in more recent studies. Finally, there were stronger reductions in income levels of individuals with chronic illness/disability in convenience samples than in random samples and in studies that did not ensure equivalence of the patient group and control group.
Although there were reduced average rates of success in individuals with chronic illness/disability for all assessed outcomes, the average effect sizes would have to be interpreted as small (Cohen, 1988). The heterogeneity of results of available studies was probably based on the fact that sufficient sample sizes are needed for identifying these effects and that success rates vary by illness characteristics.
It has been suggested that studies with nonsignificant effect sizes may often remain unpublished, thus leading to an overestimation of the effect size in meta-analyses (the "file-drawer problem"; Rosenthal, 1979). The trim-and-fill procedure analyzes the asymmetry of the effect size distribution and corrects for asymmetry that may arise from publication bias (Duval & Tweedie, 2000). While this procedure suggested that three of six effect sizes may overestimate the true effects, it also suggested that two effect sizes may underestimate the true effects. An underestimation could not be explained by a file-drawer problem. Thus, other factors may explain the asymmetry of the effect size distribution, such as whether studies with individuals with neurological diseases or sensory impairment have been conducted with large versus small samples.
Maslow et al. (2011) had suggested that chronic illness in young adults may have the largest impact on education and employment. However, the present meta-analysis indicated that the effects are largest with regard to parenthood. Four reasons may explain this finding. First, individuals often achieve other milestones (economic independence, getting married) before deciding to have children (Juang & Silbereisen, 2002). Problems with achieving other milestones would, therefore, further reduce parenthood rates. Second, fertility is reduced in some diseases, such as cancer (Dieluweit et al., 2010). Third, in the case of genetic illnesses, individuals may not want to transmit their illness to their offspring and may, therefore, decide to have no children (Klitzman, Thorne, Williamson, Chung, & Marder, 2007). Finally, as becoming parents is often the last adult milestone, many individuals have their first child in their late 20s or 30s (Juang & Silbereisen, 2002). If lower success rates indicate delayed achievement of adult milestones, low rates of parenthood may indicate that many individuals with chronic illness had not yet had time to catch up with their healthy peers. However, this meta-analysis did not find evidence of such catch-up with regard to the other assessed milestones ( Table II ).
Success rates in the field of employment and parenthood were lower in individuals with neurological illnesses, such as epilepsy and cerebral palsy, than in individuals with other physical illnesses. Impaired executive function, intelligence, and everyday competence probably explain this finding (Friedman et al., 2009). Individuals with sensory impairment reported lower success rates than individuals with other physical diseases in the field of education and employment. Higher educational facilities may not always be well adapted to the needs of students who are blind or deaf (Noble, 2010). Alternatively, lower school performance in students with sensory impairment than in individuals with other chronic physical illnesses might explain that finding. However, available studies do not support this suggestion (Pinquart & Teubert, 2012). Reduced work participation in individuals with sensory impairment may, in part, be based on the restriction of the number of available jobs to those not presupposing high levels of visual or hearing acuity. In addition, individuals with severe sensory impairment may not be able to hide their condition, which could, again, increase the risk for workplace discrimination (Chan et al., 2005).
In fact, individuals with illnesses and disabilities that are highly visible to others reported lower success with finding employment, leaving home, and marrying than their peers with invisible diseases or disabilities. However, this effect may not exclusively be explained by social discrimination because visible disabilities are often associated with restrictions in competence relevant for achieving adult milestones (Friedman et al., 2009).
A longer illness had more aversive effects on the attainment of adult milestones. This may indicate that effects of chronic illness accumulate over time. For example, longer illness and the associated earlier onset of the disease are associated, on average, with more school days missed, which could, again, negatively affect academic achievement and the development of peer relations.
Some researchers have suggested that young adults with chronic illness and disability may need more time for achievement of adult milestones than their healthy and nondisabled peers (a delayed achievement of milestones; Gledhill et al., 2000; Pacaud & Yale, 2005). However, the present meta-analysis did not find that between-group differences declined between the age of 18 and 30 years. As the study focused on the third decade of life, it was not assessed whether a catch-up might be observed in the fourth decade of life or even later. In fact, Rauck, Green, Yasui, Mertens, and Robinson (1999) observed that marriage rates tended to be lower in young adult childhood cancer survivors than in the general population, whereas no such difference was observed in adults 40 years of age and older.
The present meta-analysis found only limited empirical evidence for a decline of differences between individuals with and without chronic illness and disability in more recent studies. As about 80% of the included studies have been published after the year 2000, it was difficult to identify changes across time. Nonetheless, measures aimed at improving success with developmental milestones of adulthood—such as the enactment of disabilities acts—did not have the expected positive effects (Bell & Heitmueller, 2009).
Between-group differences in economic independence were stronger in convenience samples than in community-based probability samples. Clinical convenience samples probably include individuals with more severe levels of illness and/or disability that may restrict their level of income. Studies with nonequivalent control groups also found larger between-group differences in economic independence. The control group members of three of these studies were older than the individuals with chronic illness and disability (Laverdière et al., 2009; Miller, Spitz, Simpson, & Williams, 1982; Rikken et al., 1995), whereas control group members were younger than the patients in only one study (Bullis, Bull, Johnson, & Peters, 1995). The fact that wages tend to increase with age (Winn, 2007) could explain the larger between-group differences in studies with nonequivalent control groups. However, sociodemographic equivalence of the groups with and without chronic illness/disability did not moderate the effect sizes of the other five outcome variables.
Discussion
The present study reports results from the first broad meta-analysis on the achievement of developmental milestones of emerging/young adults with and without chronic illness and/or physical/sensory disability. Rates of success in individuals with chronic illness and disability were between 38 (parenthood) and 22% (leaving home) lower compared with their healthy and nondisabled peers. Reductions of success rates were more pronounced if individuals had a neurological disease or sensory impairment and if the disease or disability was highly visible to others. In addition, effects of chronic illness and disability on economic independence became smaller in more recent studies. Finally, there were stronger reductions in income levels of individuals with chronic illness/disability in convenience samples than in random samples and in studies that did not ensure equivalence of the patient group and control group.
Although there were reduced average rates of success in individuals with chronic illness/disability for all assessed outcomes, the average effect sizes would have to be interpreted as small (Cohen, 1988). The heterogeneity of results of available studies was probably based on the fact that sufficient sample sizes are needed for identifying these effects and that success rates vary by illness characteristics.
It has been suggested that studies with nonsignificant effect sizes may often remain unpublished, thus leading to an overestimation of the effect size in meta-analyses (the "file-drawer problem"; Rosenthal, 1979). The trim-and-fill procedure analyzes the asymmetry of the effect size distribution and corrects for asymmetry that may arise from publication bias (Duval & Tweedie, 2000). While this procedure suggested that three of six effect sizes may overestimate the true effects, it also suggested that two effect sizes may underestimate the true effects. An underestimation could not be explained by a file-drawer problem. Thus, other factors may explain the asymmetry of the effect size distribution, such as whether studies with individuals with neurological diseases or sensory impairment have been conducted with large versus small samples.
Maslow et al. (2011) had suggested that chronic illness in young adults may have the largest impact on education and employment. However, the present meta-analysis indicated that the effects are largest with regard to parenthood. Four reasons may explain this finding. First, individuals often achieve other milestones (economic independence, getting married) before deciding to have children (Juang & Silbereisen, 2002). Problems with achieving other milestones would, therefore, further reduce parenthood rates. Second, fertility is reduced in some diseases, such as cancer (Dieluweit et al., 2010). Third, in the case of genetic illnesses, individuals may not want to transmit their illness to their offspring and may, therefore, decide to have no children (Klitzman, Thorne, Williamson, Chung, & Marder, 2007). Finally, as becoming parents is often the last adult milestone, many individuals have their first child in their late 20s or 30s (Juang & Silbereisen, 2002). If lower success rates indicate delayed achievement of adult milestones, low rates of parenthood may indicate that many individuals with chronic illness had not yet had time to catch up with their healthy peers. However, this meta-analysis did not find evidence of such catch-up with regard to the other assessed milestones ( Table II ).
Success rates in the field of employment and parenthood were lower in individuals with neurological illnesses, such as epilepsy and cerebral palsy, than in individuals with other physical illnesses. Impaired executive function, intelligence, and everyday competence probably explain this finding (Friedman et al., 2009). Individuals with sensory impairment reported lower success rates than individuals with other physical diseases in the field of education and employment. Higher educational facilities may not always be well adapted to the needs of students who are blind or deaf (Noble, 2010). Alternatively, lower school performance in students with sensory impairment than in individuals with other chronic physical illnesses might explain that finding. However, available studies do not support this suggestion (Pinquart & Teubert, 2012). Reduced work participation in individuals with sensory impairment may, in part, be based on the restriction of the number of available jobs to those not presupposing high levels of visual or hearing acuity. In addition, individuals with severe sensory impairment may not be able to hide their condition, which could, again, increase the risk for workplace discrimination (Chan et al., 2005).
In fact, individuals with illnesses and disabilities that are highly visible to others reported lower success with finding employment, leaving home, and marrying than their peers with invisible diseases or disabilities. However, this effect may not exclusively be explained by social discrimination because visible disabilities are often associated with restrictions in competence relevant for achieving adult milestones (Friedman et al., 2009).
A longer illness had more aversive effects on the attainment of adult milestones. This may indicate that effects of chronic illness accumulate over time. For example, longer illness and the associated earlier onset of the disease are associated, on average, with more school days missed, which could, again, negatively affect academic achievement and the development of peer relations.
Some researchers have suggested that young adults with chronic illness and disability may need more time for achievement of adult milestones than their healthy and nondisabled peers (a delayed achievement of milestones; Gledhill et al., 2000; Pacaud & Yale, 2005). However, the present meta-analysis did not find that between-group differences declined between the age of 18 and 30 years. As the study focused on the third decade of life, it was not assessed whether a catch-up might be observed in the fourth decade of life or even later. In fact, Rauck, Green, Yasui, Mertens, and Robinson (1999) observed that marriage rates tended to be lower in young adult childhood cancer survivors than in the general population, whereas no such difference was observed in adults 40 years of age and older.
The present meta-analysis found only limited empirical evidence for a decline of differences between individuals with and without chronic illness and disability in more recent studies. As about 80% of the included studies have been published after the year 2000, it was difficult to identify changes across time. Nonetheless, measures aimed at improving success with developmental milestones of adulthood—such as the enactment of disabilities acts—did not have the expected positive effects (Bell & Heitmueller, 2009).
Between-group differences in economic independence were stronger in convenience samples than in community-based probability samples. Clinical convenience samples probably include individuals with more severe levels of illness and/or disability that may restrict their level of income. Studies with nonequivalent control groups also found larger between-group differences in economic independence. The control group members of three of these studies were older than the individuals with chronic illness and disability (Laverdière et al., 2009; Miller, Spitz, Simpson, & Williams, 1982; Rikken et al., 1995), whereas control group members were younger than the patients in only one study (Bullis, Bull, Johnson, & Peters, 1995). The fact that wages tend to increase with age (Winn, 2007) could explain the larger between-group differences in studies with nonequivalent control groups. However, sociodemographic equivalence of the groups with and without chronic illness/disability did not moderate the effect sizes of the other five outcome variables.
Source...