Children who were born in North Carolina between January 1, 2001, and December 31, 2004, and diagnosed with an OFC during their first year of life and captured by the North Carolina Birth Defects Monitoring Program (NCBDMP), were eligible for this study. The NCBDMP, an active, state-wide, population-based birth defects registry, captures births from all nonmilitary North Carolina hospitals and links vital statistics, hospital discharge, and health service use data to each infant with a birth defect (NBDPN, 2011). Children with OFCs were identified from the NCBDMP using diagnostic codes from the British Pediatric Association (749.000–749.290). Potential participants were excluded if the mothers lived outside of North Carolina, had a child with an OFC that died at any point, or if the child with OFC was adopted.

The questionnaire was developed, pilot tested, and distributed in both English and Spanish. Questionnaires were mailed between May and October 2006. Initially, the maternal residential address on the birth certificate was used to mail the surveys for all mothers of identified children with OFCs. Respondents received the survey, a recruitment letter, and a fact sheet about the study. If no response was received after 1 month, subjects were traced using publicly accessible national search databases and North Carolina health services databases. After 2 to 3 months of no response, respondents were contacted by means of telephone by trained phone interviewers. Participants who completed the survey were given a $10 gift card to a major retail store.

The survey instrument was adapted from a previous questionnaire that was designed to measure parental experiences that may impact access to care, ability to follow medical instructions, and the clinical encounter (Seid et al., 2004, 2009). The questionnaire was developed from a literature review, focus groups, and cognitive interviews with both English and Spanish speaking parents of children with chronic health conditions. The feasibility, internal consistency, and construct validity of the questionnaire were confirmed through pilot testing (Seid et al., 2004). This survey measures perceived barriers to care as multidimensional constructs on five subscales: (1) pragmatics, which included the logistical or cost related problems (9 constructs); (2) skills, which included learned strategies used to interact with the healthcare system (7 constructs); (3) marginalization, which was negative experiences within the healthcare system that parents internalize (10 constructs); (4) expectations of receiving poor quality care (6 constructs); and (5) knowledge/beliefs about popular ideas about treatment or the nature of illness that differs from mainstream medicine (four constructs) (Seid et al., 2004). We used these same five subscales in our analyses.

Thirty-nine of the total 76 open- and closed-ended survey items were specific barriers taken almost directly from the Seid et al. (2004) validated questionnaire. Thirty-five of these questions were analyzed on the five barriers to care subscales. Additional questions focused on demographic characteristics, health services use, and satisfaction of the care received. For potential perceived barriers within the five subscales, survey respondents were asked “How often were each of the following barriers a problem in the past 12 months when trying to get primary cleft or craniofacial care for your child with facial differences?” (Primary cleft or craniofacial care is the first location where receive services or the location where receive most services.) Answers were scored on a five-point Likert scale: never, almost never, sometimes, often, and almost always. Respondents could also answer not applicable. If respondents left an answer blank or marked not applicable, they were omitted from the denominator for that question only.

We also examined satisfaction with care and whether or not primary cleft and craniofacial care worked well for the child in the last 12 months in comparison with the barriers to care subscales. Due to small numbers for the question on satisfaction, we collapsed the five Likert-scale into two categories: (1) “very satisfied” and “satisfied”; and (2) “neither satisfied or dissatisfied,” “dissatisfied,” and “very dissatisfied” (Seid et al., 2004; Cassell et al., 2013).

Race and ethnicity questions were asked separately in the survey. Racial/ethnic categories included on the survey were: White, Hispanic, Black/African-American, American Indian, Alaskan Native, Asian, Pacific Islander, Native Hawaiian. We also included an open-ended “Other” category where respondents could enter their race/ethnicity, and respondents could select all that applied. Due to small numbers, we created a mutually exclusive race/ethnicity variable with categories of “non-Hispanic White” and “Other.” Thirteen respondents selected more than one race category, all 13 selected White as one of those categories and did not check Hispanic. We recoded these 13 respondents into the non-Hispanic White category (Cassell et al., 2013).

Information on health insurance coverage for both the mother and child was collected as a binary variable (yes/no). If yes, the respondents were asked to report the primary health insurance type for both the mother and child. (Primary was defined as the plan that pays the medical bills first or pays most of the medical bills.) Private insurance included enrollment in the State Employee Health Plan or any private health insurance plan purchased through an employer or directly from an insurance company. Public health insurance included the following programs: North Carolina Health Choice (state Children's Health Insurance Program), Medicaid, Carolina ACCESS, or Health Check. Military insurance included Civilian Health and Medical Program of the Uniformed Services, TRICARE, or the Veteran's Administration.

Descriptive statistics and demographic information were categorized as maternal, child, or system characteristics. Maternal characteristics included age, race/ethnicity, household income, education, marital status, number of children in the household, and number of CSHCN in the household. Child characteristics included sex, age, race/ethnicity, cleft type and the presence of other birth defects, low birth weight (<2500 grams), and preterm birth (<37 weeks). Nonisolated OFCs were categorized as an OFC diagnosis with the presence of any additional, major or minor, birth defect. If no other birth defect was present except the OFC, the OFC was categorized as isolated. System characteristics consisted of maternal and child health coverage status (yes/no) and primary insurance provider, travel time and distance to primary cleft or craniofacial center, and primary (main) language spoken in the household. One-way travel time and distance were dichotomized (≤ 60 min or >60 min and ≤60 miles or >60 miles, respectively). All data were from the survey except the child's sex, child's OFC diagnosis (cleft type and presence of other birth defects), birth weight, and gestational age, which were obtained from the birth certificate and/or the medical record. Because 24 respondents (9.8%) had missing maternal age or had illogical responses, we imputed mother's date of birth from the North Carolina vital records and calculated maternal age.

For the perceived barriers to care questions, the five-point Likert scale was collapsed and dichotomized into never/almost never (reference category) and sometimes/often/almost always due to the frequency distributions. We reclassified the sometimes/often/almost always as “ever having a problem” accessing cleft care. We also analyzed the mother's and child's insurance status by race/ethnicity, cleft phenotype, and satisfaction with cleft care to determine if there were any differences.

Bivariate analyses were conducted using chi-square and Fisher's exact tests. No multivariable analyses were conducted due to insufficient sample sizes. Data were analyzed using SAS 9.2 (SAS Inc., Cary, NC). Institutional Review Board (IRB) approvals were received from the North Carolina Division of Public Health IRB and the University of North Carolina at Chapel Hill Public Health and Nursing IRB.

Due to study limitations including the small sample size and the lack of variation in reported satisfaction of cleft care, an assessment of barriers to care and how they affect overall satisfaction could not be conducted. Our population of interest was specific to parents of children with OFCs in North Carolina, which may limit the generalizability of these results. However, the characteristics of our study sample were similar to that of mothers of children with OFCs in North Carolina overall, suggesting that our sample was representative of the population of interest. Additionally, parents may not be aware of all types of care needed for their child, which could lead to incorrect reporting of satisfaction and barriers to care. In addition, we cannot be certain of how parents interpreted “primary cleft and craniofacial care.” Parents could have interpreted this as the first place services were received or the place where most services were received. Self-reported data can introduce bias; however, studies have shown that maternal reports for child health care use are relatively accurate (D'Souza-Vazirani et al., 2005; Pless and Pless, 1995). Finally, because the survey was conducted in 2006, before the economic recession and the implementation of the Affordable Care Act, barriers to care (particularly concerning cost and logistics) and system characteristics (health insurance coverage for the mother and child) may have changed (Ghandour et al., 2014).

A strength of this study was that we used an active, state-wide, population-based birth defects registry to obtain the study population. Surveillance data allowed us to obtain and verify demographic information and access medical records to verify cleft diagnoses and presence of additional birth defects. Additionally, we used a validated barriers to care survey instrument that was developed in both English and Spanish (Seid et al., 2004). By using a validated survey, it allowed for both additional assurance that our perceived barriers were assessed appropriately and that these results potentially can be compared with barriers of care research in different populations of children with birth defects. Our study also sampled parents of children of varying ages. As pointed out by Nelson et al. (2012), most previous research focused on parent perceptions and experiences at the time of diagnosis. Our analysis examined barriers to care within the past year for children aged 2 to 6 and reported concerns that occur throughout childhood and not just after birth.

While the majority of mothers reported being satisfied with their child's cleft and craniofacial care, this questionnaire provided insight into the perceived barriers and concerns of parents of children with OFCs. Improving access and availability of services and increasing the number of facilities may minimize the time needed to obtain care for children with OFCs and may alleviate some of the most common concerns for affected families. It is also important to emphasize the need for continuous quality health insurance for families with children with OFCs and other birth defects because healthcare coverage was associated with almost 50% of all the barriers assessed in this analysis.

To the authors' knowledge, this is the first study to quantify and assess perceived barriers of care for parents of children with OFCs using a population-based, state-wide sample from a birth defects registry and a validated barriers to care questionnaire. State-wide, population-based birth defects surveillance systems provide a large base population, can allow researchers to obtain access to medical records and other health services use information, and may provide the opportunity to generalize results to other populations. Future research assessing barriers to care and identifying interventions to improve access for parents and families of children with OFCs and other birth defects could draw on both surveillance programs and validated questionnaires.