Abstract
Increasing opiate use among women of reproductive age has led to a rise in adverse pregnancy outcomes, including neonatal abstinence syndrome (NAS). Recent studies have documented the increased incidence of NAS, but subsequent impact on the chain of organizations within the social service system remains unexplored. In this article, we begin to estimate the reach of this issue by assessing the labor costs of caring for NAS infants within the Massachusetts Department of Children and Families (MA DCF). Based on a process map of services, we modeled social service hours using encounter-level hospital data as inputs. In this manner, we estimate that MA DCF professionals now devote more than 10,000 hours per month to this single problem. As opiate addiction increases across America, substantial additional investment in social service providers, foster care, Early Intervention Programs, and other family services will be required.
Background
Over the last decade, opiate abuse in the United States has increased at an alarming rate (Dart et al., 2015). Increased opiate use among women of childbearing age has brought with it an increased number of exposed newborns (Patrick, Davis, Lehmann, & Cooper, 2015; Patrick et al., 2012). Substance exposed newborns present challenges for doctors, nurses, and hospitals because detoxification can be a complex process requiring weeks of inpatient care (Kuschel, 2007).
While the rate of neonatal abstinence syndrome (NAS) admissions has quadrupled among specialized neonatal intensive care units (Tolia et al., 2015), a majority of affected infants are cared for in community hospitals (i.e., hospitals that are not characterized as specialty, teaching, or academic medical centers, normally located on rural areas of the state; Massachusetts Center for Health Information and Analysis [MA CHIA], 2015). In a 2009 national sample, the average admission costs of an infant with NAS exceeded US$50,000 (Patrick et al., 2012) and have been growing over the last few years (Patrick et al., 2015). Because NAS is highest among Medicaid recipients, the bulk of inpatient costs fall to state Medicaid programs and unreimbursed costs to hospitals.
Like Tennessee (Warren et al., 2015) and other states (Tolia et al., 2015), NAS diagnoses have increased dramatically in MA. Although much attention has been given to related hospital expenses, there are no published estimates of the costs after discharge. Infants born with opioid dependence are at high risk for subsequent mistreatment and frequently require placement in foster care (O’Donnell et al., 2009). Under “mandated reporter” statutes, these births are routinely reported to state child protection services and the infant’s family situation investigated. Ultimately, safe placement of the infant and monitoring after discharge is a complex and labor-intensive process that frequently involves the courts. In our community, we have noted a startling increase in the time spent by area social workers on this single issue. We therefore attempted to begin quantifying the statewide burden of the problem by estimating the labor costs of caring for NAS infants within the MA Department of Children and Families (DCF).
Study Data and Method
We used the Acute Hospital Case Mix data set obtained from the MA CHIA. The case mix data set contains encounter-level data concerning all inpatient and outpatient hospital visits in the state. Data are submitted quarterly by all acute care hospitals and compiled by fiscal year (FY) according to strict quality standards set by CHIA. Among many other elements, case mix data contain residence zip codes, medical diagnoses, lengths of stay, and hospital charges for all individuals admitted to MA’s hospitals. For this analysis, we used deidentified inpatient admission data from 2004 to 2013. Our work was reviewed and approved by the CHIA Data Release Board and administratively approved by the institutional review board at Boston Children’s Hospital.
We identified infants with NAS through International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 779.5 (Drug Withdrawal Syndrome in Newborn). From this initial cohort for the decade (N = 6,976), we excluded infants who were transferred from another acute hospital (n = 662) to avoid double counting and also infants with an out-of-state permanent zip code of residence (n = 75). Infants with presumed iatrogenic drug withdrawal were also excluded following the criteria discussed in Patrick et al. (2012), and the remaining babies (n = 6,170) were then assigned to one of the four DCF regions based on their zip code of residence.
To estimate the immediate impact that NAS infant care had on the child welfare system, we interviewed supervisors and caseworkers within the MA DCF. Based on these interviews, we created a process map of services that routinely flow from DCF to all substance-dependent newborns (Figure 1). When multiple paths were possible, we estimated probabilities according to historical experience. Briefly, all infants required a home assessment and series of family, medical, and legal interviews to generate a placement decision. With some regional variation, historical experience teaches that close to half of these investigations result in home placement (usually addicted mothers in treatment) and half in custody placement. The initial placement phase can be highly variable, requiring a minimum of approximately 10, a maximum of 24, and a mode of 14 work hours. Of infants in home placement, monitoring continues for 45 working days, after which one quarter of cases are typically closed and three quarters remain open. Of infants requiring custody placement, one quarter typically enter foster care and three quarters enter some form of kinship care. As depicted in the Figure 1, the time requirements of each path are different. After placement, all infants are followed by a social worker in an ongoing care process requiring coordination with the courts, visitation, medical follow-up, and administrative work including a written comprehensive assessment and service plan in the third month.
Department of Children and Families process map.
After vetting the process map and approximate work hours with the area directors of the most affected districts, we used birthing data as inputs to a process-based model and estimated the total number of work hours required each month to investigate, place, and provide follow-up supervision for all opioid dependent infants to 6 months of age. Because infants are assigned to DCF areas by residence, we used zip codes to apportion work hours to regions and area offices. Trends were estimated using linear regression and all analyses were performed using Python, an open-source programming language (Python Software Foundation, 2015).
Results
Statewide, opioid dependent infant births increased sixfold over the period (Figure 2). Although the average length of hospital stay decreased from 21.2 to 19.3 days (p < .01), median charges increased from US$20,981 to US$30,043 (inflation adjusted to 2015 dollars, p < .01, see Table 1). Infants were predominantly born to white Medicaid recipients with a lower number of affected Hispanic, Asian, and African American infants remaining stable over the period. Across all years, 63–65% of infants were cared for in community hospitals.
Evolution of the number and rate of neonatal abstinence syndrome in Massachusetts from 2004 to 2013.
Demographics and Characteristics of NAS Hospital Stays.
Note. NAS = neonatal abstinence syndrome; FY = fiscal year; DCF = Department of Children and Families; CI = confidence interval.
Using an estimated follow-up period of 6 months, the cumulative monthly number of infants followed by the DCF rose dramatically. In the period between October 2010 and August 2013, the monthly number of infants in the DCF system increased from 273 to 357. Notably, the problem in MA is suburban and rural, with the city of Boston experiencing lower numbers than other regions.
The estimated number of hours devoted to the care and protection of NAS infants is depicted in Figure 3 for the same period. There was considerable variability among the state’s 29 area offices, with some devoting an estimated 800 hr per month or more to this problem alone.
Monthly Department of Children and Families region work hours for neonatal abstinence syndrome infants for the period between October 2010 and September 2013.
Discussion
The opiate addiction epidemic in MA has resulted in a more than sixfold increase in the number of infants born with opiate dependence between 2004 and 2013. As mandated reporters, hospitals routinely report all substance-exposed newborns to the DCF, which has the responsibility of ensuring their safe placement, health, and well-being. Here we show that, with large regional variation, the growing burden of this problem accounts for a substantial expenditure of social service resources. Statewide, we estimate that it accounted for a monthly average of about 10,650 hr in the 2013 FY. In the most affected area offices, this single issue occupied the equivalent of about 800 monthly hours. To estimate the financial burden, we use the average hourly wage of a DCF case manager/social worker as of August 2015, US$33.35. 1 While there is a small regional variation in average salaries, we approximate the salaries in all regions by this number and estimate that, in the 2013 FY, this problem consumed around 4.3 million dollars in labor costs alone. To put this into perspective, the total budget for case managers and social workers for the same year was 170.6 million dollars (inflation adjusted; MA Budget and Policy Center, 2015), meaning that our estimates indicate that about 2.5% of the total state budget for personnel was spent on this single problem.
We suspect that the regional variation in neonatal exposure simply reflects the regional differences in opioid abuse more generally. As reported elsewhere, the opiate epidemic is an increasingly suburban problem (Brown, 2013; Cicero, Ellis, Surratt, & Kurtz, 2014), and we sought to confirm this by calculating the total number of hospitalizations for opioid abuse within each DCF area and zip code since 2010. As expected, we found similar regional and temporal patterns with dramatic increases in suburban hospitalizations while those in Boston remained stable. A deeper understanding of these regional differences will require further investigation but could have extremely significant policy implications.
There are clear limitations to this work. First, identification of opioid dependent infants by ICD-9 requires accurate coding and may underestimate the actual incidence (Burns & Mattick, 2007). For example, some infants born to substance abusing mothers may not exhibit symptoms while in the hospital or may not return for hospital care and can therefore be missed (Smirk, Bowman, Doyle, & Kamlin, 2014). We believe that overestimates are unlikely, however, since the NAS diagnosis is straightforward, carries important reporting responsibilities, and the infants coded in our data shared long hospital stays. Second, the work hours we assigned are based on interviews from a sample of social workers whose recollections and estimates may be inaccurate. We attempted to improve our estimates by breaking down tasks, creating and refining a process map, vetting it widely, and corroborating the hours spent on each task with frontline providers and their supervisors. Although we believe these estimates to be reasonable, we cannot claim precision. Third, estimates here are not generalizable beyond MA, since state policies, procedures, and practices can vary significantly. Even so, we believe that the techniques and tools employed here can easily be applied elsewhere to connect public health problems to social service costs. Fourth, we neglect the additional, often substantial, burdens upon attorneys, judges, and other members of the criminal justice system (National Center on Addiction and Substance Abuse, 2009). We considered this to be beyond our original scope and a separate subject of investigation. Finally, we arbitrarily modeled only the first 6 months of each infant’s life. In reality, these infants typically remain in the social service system much longer and carry significant additional costs (Derrington, 2012). Beyond 6 months, however, the wide range of possible care paths, and uncertainty around them, made modeling impractical.
In 2005, MA devoted 21.8% of its total state budget to substance abuse and addiction (National Center on Addiction and Substance Abuse, 2009). The opiate problem has since increased dramatically during a period of economic recession. Beginning in 2008, diminishing tax revenues in MA forced sharp reductions in DCF funding with child welfare expenditures ultimately falling below 2001 levels. Estimates offered here suggest that very sharp increases in social service providers, drug treatment/prevention centers, and Early Intervention Programs will be necessary as the economy recovers.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.