Awareness and Use of California's Paid Family Leave Insurance Among Parents of Chronically Ill Children FREE

Chronically ill children, or children with special health care needs, constitute 13% to 17% of children in the United States.^1- 3 Children with special health care needs average 3 times as many medical encounters as other children,^1- 4 account for one-half of child hospital days,¹ and miss nearly 3 times as much school.⁵ Their health-related needs create substantial pressure on parents to miss work.^6- 7

Parents of children with special health care needs were a major target of the 1993 Federal Family and Medical Leave Act (FMLA), which guarantees eligible workers up to 12 weeks of unpaid leave with job protection (ie, protection from being fired) to care for themselves or ill family members.⁸ However, only 47% of employees are eligible for FMLA leave,⁹ and many cannot afford unpaid leave.⁸ Of the 3.5 million employees who needed but could not take leave in 2000, 78% cited inability to afford leave as a reason; of these, 88% said they would have taken leave had they received some pay or additional pay.⁸ Availability of employer-provided leave, meanwhile, ranges widely.¹⁰ Many employees use sick leave (often violating employer policies) or vacation to care for family; other employees have no access to leave.¹¹

Consequently, several states have passed paid family leave legislation,^12- 15 with additional states and Congress exploring it.^16- 20 Our prior research suggests that paid leave might be important for parents of children with special health care needs—parents with access to employer-provided paid leave were less likely to have unmet need for leave,⁷ and paid leave was associated with parent-perceived positive effects of leave on their emotional health and finances and their children's health.²¹ California was the first state to offer paid family leave.¹² In 2004, its Paid Family Leave Insurance Program (PFLI) began providing 6 weeks of non–job-protected paid leave annually for most part-time and full-time employees at approximately 55% of salary (maximum 2008 benefit, $917 per week).¹² Benefits apply after employees miss 1 week of work (continuously or cumulatively). A mandatory payroll tax (currently approximately $1 per week, subject to increase if PFLI expenditures grow) is added to the State Disability Insurance wage-statement line item and reserved in a statewide pool. A statement signed by a physician documenting the illness is required.

The Paid Family Leave Insurance Program has been a model for state and federal paid family leave efforts, yet its impact is unknown. We surveyed 2 successive cohorts of employed parents of children with special health care needs. We examined parents’ reports of taking leave and need for leave before PFLI, PFLI awareness and use approximately 18 months after benefits began, and PFLI's effect on taking leave.

We sampled children receiving inpatient or outpatient care at 2 large tertiary-care referral centers (both major referral centers for children with special health care needs): Children's Memorial Hospital, Chicago, Illinois (abbreviated as Illinois) and Mattel Children's Hospital, University of California, Los Angeles (abbreviated as California). Children's hospitals like Children's Memorial Hospital and Mattel Children's Hospital provide the majority of highly specialized care for children with complex and rare conditions.²² Because Illinois does not have a paid leave program, Children's Memorial Hospital served as a comparison. Both states have similar labor markets, particularly regarding unemployment,^23- 24 average per capita income,^25- 26 and nonfarming job distribution.²⁷

We identified children with special health care needs by adapting a validated International Classification of Diseases, Ninth Revision billing code approach.^28- 31 The unmodified approach generally yields samples dominated by common diagnoses that typically require few missed work days. Because we were most interested in children with special health care needs whose parents needed to miss work, we restricted the code-list to disease categories⁷ with the highest average per-patient physician charges, which have been shown to correlate with illness severity.^4,32- 34 These disease categories included bronchopulmonary dysplasia, cerebral palsy, chronic anemias, chronic enteritis/colitis, chronic renal failure, congenital heart diseases, cystic fibrosis, degenerative neurological disorders, hydrocephalus, immunologic disorders, malignancies, organ transplant complications, and rheumatologic disorders. We omitted human immunodeficiency virus at institutional review board request.

Using this list, we recruited 2 successive cohorts of children by identifying all children (<18 years) in each hospital's billing databases who were assigned a qualifying diagnosis between October 1, 2002, and September 30, 2003 (before PFLI became available) and between October 1, 2004, and September 30, 2005 (after PFLI), listed as alive, and living in their respective states.

Because PFLI is available only to employed individuals, we focused on children with special health care needs with employed parents. However, the hospital databases that we used did not include any information regarding parent employment status. To undersample parents who were less likely to be employed, we stratified by Medicaid status, the best proxy for employment status available in the databases (details appear elsewhere⁷).

The study was designed to detect a reduction from an anticipated 40% of parents at the California site with unmet need for leave before PFLI to 25% after (80% power, 2-sided P < .05 for difference-of-differences analysis), with no corresponding change from a 40% Illinois baseline. This required 380 completes per site per wave among employed parents (designed to be 70% of our sample). We projected a 70% response rate and a design effect of 1.15 from sample design/nonresponse weighting, requiring an initial sample size of 777 per site per wave.

We identified children with special health care needs at the California site (1570 before and 1499 after PFLI) and at the Illinois site (3680 before and 4752 after PFLI). At each wave, we randomly selected separate cohorts of 800 children from each site.

Telephone interviewers conducted 40-minute computer-assisted telephone interviews (in English and Spanish). Our protocol for sampling parents in 2-parent households appears elsewhere.⁷

Before PFLI benefits began (November 21, 2003-January 31, 2004), we interviewed 562 parents at the California site and 554 parents at the Illinois site. Excluding parents who were never located (11%) or who were otherwise ineligible (2% had moved out-of-state or their child had died), response rates were 82% at the California site and 80% at the Illinois site.³⁵ Approximately 18 months after benefits began (November 18, 2005-January 31, 2006), 583 parents at the California site and 512 parents at the Illinois site participated. Response rates were 81% at the California site and 74% at the Illinois site.

We received Health Insurance Portability and Accountability Act waivers to use hospital databases to obtain necessary private health information. The protocol was approved by all participating institutional review boards.

Leave Usage. At each wave, we asked parents of children with special health care needs how many days of work they had missed in the past year because of their child's illness. We also asked parents about unmet need for leave (“Was there any time in the last 12 months that you felt you needed to miss work because of [child]'s illness, but you did not miss work?”). Parents who said yes were asked the reasons why they did not take leave (>1 answer could be selected). Parents were also asked about their longest leave ever due to their child's illness, whether they had cut short that leave, and why.

Awareness and Use of PFLI. After PFLI began, we asked parents at the California site whether they had heard of and used PFLI. Parents who had heard of but not used PFLI were asked why not (>1 answer could be selected).

We collected standard demographic and employment data (Table 1 and Table 2). Some items came from the US Department of Labor Survey of Employees.⁸ Other items were developed by the researchers; reviewed by clinicians, attorneys, and social scientists familiar with children with special health care needs and labor issues; and piloted among parents of children with special health care needs. Parents were asked about self-reported race/ethnicity (using investigator-defined categories) because of its relevance to family leave studies.¹¹ We included overtime eligibility because having a nonsalaried position that is eligible for overtime often indicates a less-flexible work schedule.

We collected 3 child illness-related measures: (1) a PedsQL short version,³⁷ (2) number of hospitalizations in past year, and (3) number of hospital nights in past year. The PedsQL, a standard measure of pediatric health-related quality of life, does not exist for children younger than 2 years; we assigned these children the mean PedsQL score in our data and marked them with an indicator variable, allowing them to be included in models without biasing estimates of PedsQL-related associations. The number of hospitalizations and hospital nights allowed us to examine whether experiencing serious illness episodes that required hospitalization was associated with PFLI awareness and use.

Analyses incorporated weights accounting for nonresponse (inverse predicted probabilities of multivariate logistic regressions) and Medicaid-status stratification. No variable was missing more than 3% of observations before PFLI began or more than 4% after PFLI. To prevent bias from limiting multivariate analyses to complete cases,³⁸ we used a single-chained equations imputation.³⁹

The analysis sample for both waves consisted of all part-time and full-time employed parents (372 from California and 382 from Illinois before PFLI began, and 396 from California and 370 from Illinois after PFLI). To analyze whether PFLI resulted in changes in taking leave and unmet need for leave, we used a difference-of-differences approach, which compared change at the California site (with PFLI) with change at the Illinois site (without PFLI). To analyze PFLI awareness and use, we used them as outcome variables in bivariate logistic regressions for parents at the California site surveyed after PFLI, with Table 1 and Table 2 variables as predictors. Although too few parents (n = 20) had used PFLI to conduct multivariate analyses of use, we were able to conduct multivariate logistic regression of awareness, using a limited number of predictors (those with P < .05 bivariately) to avoid overfitting.^40- 41 Finally, we performed bivariate subanalyses among parents who were aware of PFLI to examine predictors of use.

Analyses were performed by using Stata version 10 (StataCorp, College Station, Texas). Reported P values were 2-tailed and significance level was set at P < .05.

Respondent study flow is shown in Figure 1 and Figure 2. Table 1 and Table 2 show characteristics of California and Illinois respondents.

California's PFLI did not increase the percentage of parents taking leave from before to after initiation of the program (Table 3). Before PFLI began, 295 parents (81%) at the California site and 290 parents (78%) at the Illinois site took at least 1 day of leave in the previous year to care for their ill child compared with after PFLI began (327 parents [79%] at the California site and 296 parents [79%] at the Illinois site). In a difference-of-differences analysis comparing change at the California site (with PFLI) with change at the Illinois site (without PFLI), California reported no change relative to Illinois in taking at least 1 day of leave (difference-of-differences analysis, −3%; 95% confidence interval [CI], −13% to 7%).

The Paid Family Leave Insurance Program also did not increase the amount of leave parents took (Table 3). Before PFLI began, 64 parents (21%) at the California site and 51 parents (14%) at the Illinois site took at least 4 weeks of leave compared with after (74 parents [19%] at the California site and 38 parents [11%] at the Illinois site). In a difference-of-differences analysis, California reported no change relative to Illinois (difference-of-differences analysis, 1%; 95% CI, −9% to 10%).

Before PFLI began, 152 parents (41%) at the California site and 138 parents (36%) at the Illinois site, and after PFLI began, 156 parents (41%) at the California site and 135 parents (37%) at the Illinois site said that, at least once in the past year, they did not miss work despite believing their child's illness necessitated it (Table 3). In a difference-of-differences analysis, California reported no change relative to Illinois (−1%; 95% CI, −13% to 10%).

Before PFLI began, the most-cited reasons (California and Illinois combined) for not missing work were could not afford to miss income (179 [24%]), might lose job or business (118 [17%]), might hurt career (108 [15%]), work was too important (103 [13%]), could not afford to miss benefits for child (111 [13%]) or self (97 [11%]), and did not want to use up time off (91 [11%]). Of parents whose longest leave was unpaid, 141 (68%) said they cut it short for financial reasons.

Main Analyses . Only 77 parents (18%) reported having heard of PFLI and only 20 (5%) reported using it (Table 4 and Table 5). In bivariate analyses, overtime-eligible parents (generally with less flexible schedules) compared with ineligible parents were more likely to have heard of PFLI (45 [29%; 95% CI, 21%-39%] vs 32 [10%; 95% CI, 7%-14%]; P < .001) and used PFLI (13 [9%; 95% CI, 5%-18%] vs 7 [2%; 95% CI, 1%-4%]; P = .004). Parents who were employed full-time (63 [20%; 95% CI, 15%-27%]) compared with part-time (14 [10%; 95% CI, 6%-17%]; P = .03) were more likely to have heard of PFLI, but were not more likely to have used it. Compared with parents with full-time–employed spouses (43 [20%; 95% CI, 14%-28%]), single parents were less likely to have heard of PFLI (12 [9%; 95% CI, 5%-15%]; P = .02), but not less likely to have used it. Use of PFLI also increased with the number of nights spent in the hospital (odds ratio [OR], 1.07 for each additional week in the hospital; 95% CI, 1.00-1.15). Finally, neither awareness nor use of PFLI was significantly associated with parent education, access to employer-provided paid leave, household income, child PedsQL score, or number of hospitalizations in the past year.

In a multivariate analysis predicting PFLI awareness, overtime-eligible parents were more likely to have heard of PFLI than ineligible parents (OR, 0.31; 95% CI, 0.17-0.57). No other significant bivariate associations remained significant in multivariate analysis.

Subanalyses. Even among subsets of parents with greater potential need for paid leave, PFLI awareness and use were minimal. Among parents who, despite having no access to paid leave beyond sick, vacation, or other non-PFLI leave, still missed work in the past year due to their child's illness, 102 (71%) missed enough days (>5) to qualify for PFLI; however, only 30 (15%) reported having heard of it and 10 (6%) used it. Among parents who reported unmet need for leave, 20 (9%) had heard of PFLI and 5 (2%) had used it.

Before PFLI, 63 parents (22%) at the California site and 57 parents (15%) at the Illinois site had no access to paid sick or vacation leave; after PFLI, 70 parents (26%) at the California site and 54 parents (16%) at the Illinois site had no access to paid sick or vacation leave. In addition, 219 parents (74%) at the California site and 247 parents (78%) at the Illinois site had no access to other non-PFLI paid leave before PFLI began, and 248 parents (77%) at the California site and 245 parents (81%) at the Illinois site had no such access after PFLI. In a subanalysis among parents who had heard of PFLI, parents who did not have access to vacation, sick, or other non-PFLI paid leave were more likely to have used PFLI than those with such access (3 [63%; 95% CI, 27%-89%] vs 16 [20%; 95% CI, 12%-32%]; P = .01).

Among parents who had heard of PFLI but had not used it, 34 (59%) used vacation, sick, or other non-PFLI paid leave instead, 21 (41%) did not know about or understand the program at the time, 8 (11%) missed less than the required waiting period, and 3 (10%) said their supervisors told them not to take time off even though they were eligible for PFLI (Table 6).

Twenty-eight parents (6%) reported that a physician or other hospital/office staff person had discussed leave laws with them. Those parents reported that the topic was raised in these discussions by the child's physician (6 [28%]), other office/hospital staff (14 [50%]), the parent (7 [19%]), or another family member (1 [3%]).

Initial implementation of the nation's first paid family leave law failed to reach a large and vulnerable target constituency: parents of children with special health care needs. Approximately 18 months after PFLI began, only 18% of employed parents in our sample—drawn from a facility that serves the sickest of the sick²²—had heard of PFLI and only 5% had used it. Even especially vulnerable subgroups (eg, parents without employer-provided paid leave, parents with unmet need for leave) reported low PFLI awareness and use. Results are particularly surprising because, although we intentionally studied children who were sicker than children with special health care needs nationally,^22,42 their parents' employment status still matched national trends,⁴³ making them especially likely to need leave.

Many factors may explain the minimal use of PFLI, but lack of awareness is likely important. Uptake of new policies generally requires a combination of awareness, low perceived costs (eg, minimal income loss), and high perceived benefits (eg, improving children's health, allaying children's fears).^44- 46 Currently, PFLI requires that employers provide a pamphlet only to new employees (along with other benefits descriptions) and employees who take leave for a covered purpose,⁴⁷ meaning that some employees must know about PFLI before asking about it or decide to take leave before knowing that they could receive pay. Indeed, many employees might never apply because they presume leave would be unpaid. Even wage statements collapse the PFLI tax into the State Disability Insurance line, which (perhaps inadvertently) further hides the existence of the benefit from employees.

In contrast with PFLI, FMLA was accompanied by a 2-year Department of Labor publicity campaign and mandatory posting of FMLA information at workplaces. In 1995, at the end of the campaign, 58% of employees who worked for employers affected by FMLA knew about the law.⁴⁸ This number stands in marked contrast to the 18% of employed parents in our sample who were aware of PFLI. The discrepancy seems especially striking given that our population had substantial reason to be aware of leave options.

Potential barriers to uptake, however, do not end with lack of awareness. The PFLI does not protect employees from being fired, although employees who are also covered by FMLA are protected. Among parents in our sample who had heard of PFLI but not used it, reasons included supervisors telling them not to take time off even though they were eligible (10%) and fear of being demoted or fired (7%). An explicit job protection provision might help parents feel more confident in using PFLI.

Another potential barrier is PFLI's requirement that employees miss more than 1 week of work before receiving benefits. Among parents in our sample who had heard of but not used PFLI, 11% reported not using it because they missed less than the mandatory waiting period.

Finally, PFLI's provision of partial pay (approximately 55% of one's salary up to a cap) may have limited its use. In our prior work, we found that, although parents who received full pay during leave were more likely than parents who received no pay to report positive effects of leave on health and finances, the same was not true for parents who received only partial pay.²¹ If perceived benefits of paid leave diminish at lower rates of pay, 55% may not be enough to persuade some parents to apply.

This lack of awareness and use are concerning, because our data indicate a substantial need for leave among parents of children with special health care needs. Seventy-nine percent of baseline parents at California and Illinois sites took at least 1 day of leave to care for their child; 18% took at least 4 weeks. Nevertheless, 38% of parents reported not missing work when they needed to for their child's illness, often citing an inability to afford missed income. Likewise, among parents whose longest leave was unpaid, 68% said they cut it short for financial reasons.

Our study has limitations. Our sample, which was limited to 2 institutions in 2 states, may not generalize to all parents of children with special health care needs. A larger study might be more broadly representative and might find variations by geographic region, job type, and child illness category. In particular, our diagnosis list did not include serious mental health conditions and, although we included the number of hospital admissions and hospital nights as predictors, our condition selection did not directly incorporate severity grading. Furthermore, it is unclear to what degree lack of awareness of PFLI reflected a lack of effort by employers and the state to disseminate information vs poor parental recall. Parents might have learned about PFLI but then forgot about it. If this were the case, however, it would still be concerning, because our population is one for which this information would likely be especially salient. Such a scenario would simply reinforce questions about the effectiveness of PFLI dissemination.

Although our findings are generally disappointing, they are not without promise. In our subanalysis of parents who had heard of PFLI, parents were more likely to have used it if they did not have access to sick, vacation, or other non-PFLI paid leave. This suggests that, among parents who are aware of the law and undeterred by its barriers to use, PFLI might be a viable option for those who lack employer-provided leave or who need more leave than their employer provides. Forty-three percent of US private-sector workers have no paid sick leave, 33% have no paid vacation, and 62% have no paid personal leave.⁴⁹ Thus, if lack of awareness and other barriers can be reduced, there is reason to believe that a substantially larger pool of parents will benefit from having PFLI.

Almost no parents reported that clinicians had discussed leave laws with them; however, clinicians (including health educators and social workers) can play a valuable role in informing parents about leave and leave laws. The American Academy of Pediatrics emphasizes parents' importance in promoting the health of ill children and advocates that clinicians adopt a family-centered approach.^50- 51 When children are hospitalized or diagnosed with chronic disease, it may be a particularly relevant moment to teach parents about family leave options and to discuss strategies for balancing job demands and child needs. Clinicians can advise parents of children with special health care needs that leave should be considered in family health and employment decisions, and that taking leave might improve child health and parent emotional health.

For policymakers considering paid leave programs, our findings suggest that it is insufficient for employees to learn about the program only when starting a new job or requesting leave. Additional dissemination (eg, media campaigns, periodic employer-based notification of all employees) may raise awareness. Maximizing uptake of paid leave programs among parents of children with special health care needs may be a particularly important policy goal, given their substantial need for leave.

Although our findings are preliminary, they highlight both the need for paid leave options among parents of children with special health care needs in our sample and the initial failure of California's PFLI program to meet that need. Before PFLI's implementation, many employed parents of children with special health care needs receiving care at the hospitals in our study experienced financial constraints that precluded taking leave or required them to cut short their leave. Although PFLI was designed to minimize a leave's financial hardship, parents in our California sample were generally unaware of PFLI and rarely used it, and PFLI neither increased the frequency or duration of taking leave nor decreased unmet need for leave compared with parents in the Illinois sample. Dissemination strategies may need examination before individual states or the United States as a whole adopt similar policies.