129x Filetype PDF File size 0.30 MB Source: scholar.harvard.edu
ORIGINAL ARTICLE Out-of-Pocket Medical Costs and Third-Party Healthcare Costs for Children With Down Syndrome 1 1 1 2 3 Andrew Kageleiry, David Samuelson, Mei Sheng Duh, Patrick Lefebvre, John Campbell, and Brian G. Skotko4,5* 1 Analysis Group, Inc., Boston, Massachusetts 2 Groupe d’analyse, Ltee, Montreal, Quebec, Canada 3 Department of Economics, Harvard University, Cambridge, Massachusetts 4 Division of Medical Genetics, Department of Pediatrics, Massachusetts General Hospital, Boston, Massachusetts 5 Harvard Medical School, Boston, Massachusetts Manuscript Received: 24 June 2016; Manuscript Accepted: 27 October 2016 Prioranalyseshaveestimatedthelifetimetotalsocietalcostsofa person with Downsyndrome(DS);however,nostudiescapture How to Cite this Article: the expected medical costs that patients with DS can expect to Kageleiry A, Samuelson D, Duh MS, incur during childhood. The study utilized the OptumHealth Lefebvre P, Campbell J, Skotko BG. 2016. Reporting and Insights administrative claims database from Out-of-pocket medical costs and 1999 to 2013. Children with a diagnosis of DS were identified, third-party healthcare costs for children andtheirtimewasdividedintoclinicallyrelevantagecategories. with Down syndrome. Patients with DS in each age category were matched to controls AmJMedGenetPart A 9999A:1–11. without chromosomal conditions. Out-of-pocket medical costs and third-party expenditures were compared between the pa- tient-age cohorts with DS and matched controls. Patients with DS had significantly higher mean annual out-of-pocket costs information [Skotko, 2005]. Among the many questions asked: than their matched controls within each age and cost category. HowexpensiveisittoraiseachildwithDS?Canourfamilyafford Totalannualincrementalout-of-pocketcostsassociatedwithDS to pay for the medical expenses? What can wereasonably expect to were highest among individuals from birth to age 1 ($1,907, be covered by our medical insurance provider? While the decision P<0.001). The main drivers of the incremental out-of-pocket to continue or terminate a pregnancy often involves complex costs associated with DS were inpatient costs in the 1st year of considerations, accurate and up-to-date information on health life ($925, P<0.001)andoutpatientcostsinlateryears(ranging care expenditures is an important element for some families. $183–$623, all P<0.001). Overall, patients with DS incurred Since October 2011, noninvasive prenatal screening (NIPS), incremental out-of-pocket medical costs of $18,248 between utilizing plasma cell-free DNA, has enabled expectant couples to birth and age 18 years; third-party payers incurred incremental learn as early as 9 weeks into gestation whether their fetus might costs of $230,043 during the same period. Across all age catego- haveDS[Bianchietal.,2012;Nicolaidesetal.,2012;Nortonetal., ries, mean total out-of-pocket annual costs were greater for 2012;Zimmermannetal.,2012].Asresearchcontinues,marketing individualswithDSthanthoseofmatchedcontrols.Onaverage, widens, and insurance coverage expands, more expectant women parentsofchildrenwithDSpayanadditional$84permonthfor will have access to NIPS, inevitably leading to a greater number of out-of-pocket medical expenses when costs are amortized over couples receiving a prenatal diagnosis of DS. 18 years. 2016 Wiley Periodicals, Inc. Key words: Down syndrome; children; out-of-pocket costs; Andrew Kageleiry and David Samuelson are co-first authors. Abbreviations: CPI, Consumer Price Index; DS, Down syndrome; HMO, third-party payers Health maintenance organization; NIPS, Noninvasive prenatal screening; POS, Point of service; PPO, Preferred provider organization. Correspondence to: Brian Skotko, M.D., M.P.P., Massachusetts General Hospital, 185 INTRODUCTION Cambridge Street, Room 2222, Boston, MA 02114. E-mail: bskotko@mgh.harvard.edu With prenatal testing for Down syndrome (DS) rapidly evolving, Article first published online in Wiley Online Library an increasing number of expectant couples must make decisions (wileyonlinelibrary.com): 00 Month 2016 about their pregnancy options, often with limited time and DOI 10.1002/ajmg.a.38050 2016 Wiley Periodicals, Inc. 1 2 AMERICAN JOURNAL OF MEDICAL GENETICS PART A Researchers have madeargumentsthatprenataltestingforDSis Patient Selection costeffective,andtheircalculationsgenerallyincludeanincremental Patients with at least one medical claim associated with a diagnosis lifetime cost of raising a child with DS (i.e., the excess lifetime cost of DS (ICD-9-CM code: 758.0x) were eligible for inclusion in the comparedtoatypically developing child) [Ball et al., 2007; Cuckle study cohort. Patients were also required to be enrolled in their et al., 2013; Song et al., 2013]. Ball et al. [2007] concluded that the family insurance plan as a child (i.e., <18 years of age) and to have incremental“societalcostofraisingandcaringforanindividualwith an identifiable mother and/or father on the insurance plan (for- DS is $762,748,” adjusted to 2006 USD. Song et al. [2013] more mally designated as the “plan subscriber” or “spouse of the recently stated that the incremental “cost of Down syndrome” was subscriber”). Patients were further required to have discernible $677,000in2012USD,witharangethatcouldbeashighas$800,000 demographic,enrollment, and parental characteristics which were over a lifetime. Cuckle et al. [2013] price the lifetime incremental used in a later matching algorithm. costs around $900,000 in 2013 USD. All of these papers, however, Individuals withoutanydiagnosesforchromosomalconditions draw their data from the same primary source where incremental (ICD-9-CMcode:758.xx)intheirobservable medical claims were direct costs (e.g., extra expenses from inpatient hospital stays, eligible for inclusion as potential controls. These patients were also outpatient medical visits, long-term care, developmental services, required to be enrolled as children, have an identifiable mother andspecialeducation)andindirectcosts(e.g.,lostproductivitydue and/orfather,andhavethesameaforementionedcharacteristicsto to morbidity and early mortality) were estimated for “the entire be used in the matching algorithm. lifespan”forpersonswithDSborninCaliforniain1988[Waitzman The current study was conducted to estimate the incremental et al., 1994]. This study is now dated and demographically non- yearly medical and pharmacy out-of-pocket costs incurred by applicabletomanypartsoftheUnitedStates.Perhapsmostimpor- patients with DS relative to controls without DS across a number tantly, these studies estimated societal costs (relevant for ofdifferentdevelopmentaltimeperiods.Accordingly,theperiodof cost-effectivenessandcost-benefitstudies)andnotindividualcosts, observationduringwhichpatientswereenrolledintheirinsurance which are likely of greater interest to expectant parents. plan was split into clinically relevant age categories. Following the Few studies have estimated such costs. According to the AmericanAcademyofPediatricshealthcareguidelinesforpatients 2009–2010 National Survey of Children with Special Healthcare withDS,thefollowingtimeperiodswerestudied:<1yearold,1to Needs, approximately 23.6% of families who have a child with DS <3yearsold, 3 to <5 years old, 5 to <13 years old, and 13 to <18 reported that they pay $1,000 or more in out-of-pocket medical yearsold[Bull,2011].Afterselection,patientsinboththestudyand expenses each year [Child and Adolescent Health Measurement potential control cohorts were assigned to one or more age Initiative,2009/2010].About32%ofthesefamiliesreportedthatDS categories depending on the period during which they were “caused financial problems for the family” [Child and Adolescent enrolled in their insurance plan. For individuals with more than HealthMeasurementInitiative,2010/2009].Basedonthepaucityof one continuous enrollment period, only the first was selected, such information in the existing literature, the primary purpose of regardless of length; patients who were continuously enrolled this study is to use a large database to estimate the incremental across multiple age groups were assigned to each of the respective out-of-pocket medical costs for patients with DS, between birth to groups (Fig. 1). 18yearsofage,intheUnitedStates.Wealsocalculatethetotalhealth care expenditures for private third-party payers in order to provide Matching Algorithm updated data for policymakers and other key stakeholders. Tocontrol for observable confounding factors, a greedy matching algorithm was used to match, in a one-to-four ratio, individuals MATERIALS AND METHODS withineachobservedagecategoryinthestudycohort(“patient-age Data observations”) to patients within the same age group in the potential control [Miettinen, 1969; Bergstralh and Kosanke, This retrospective matched-cohort study was conducted utilizing 1995, 2004]. Greedy matching randomly locates study-control data from the OptumHealth Reporting and Insights database. The pairsthatmeetapre-specifiedsetofcriteria;thismethodisdistinct data are based on administrative claims for privately insured indi- fromoptimalmatchingalgorithmsinthatitmaximizessamplesize viduals covered by 82 self-insured Fortune 500 companies with while also reducing selection bias [Parsons, 2001]. Patient-age operations in a broad array of industries and job classifications observationswerematchedexactlyongender,lengthofcontinuous across the United States. The OptumHealth database eligibility within an age group (180 days), the calendar year contains medical claims (dates of service, diagnoses received, pro- (exact), and their age (1 year) at the beginning of continuous cedures performed, places of service, and payment amounts), eligibility withinanagegroupandtheaverageageoftheirparent(s) outpatient pharmacy claims (fill dates, national drug codes, and at the time of their birth (1 year). Such parental characteristics payment amounts), and eligibility information (patient demo- wereincludedinthematchingalgorithmtocontrolfordifferences graphics and enrollment history) for over 18 million individuals, in factors that may affect a parent’s choice for his or her child’s including primary subscribers as well as their covered beneficiaries medical care. Additionally, an unconditional logistic regression (i.e., spouses and dependents).Thedatabasecoverstheperiodfrom propensity score was estimated for each patient-age observation, Q11999toQ12013andhasbeencitedfrequentlyinnumerouspeer- modeling the likelihood of a patient-age observation being in the reviewed studies of medical costs [Loftus et al., 2014; Rice et al., DS cohort. Controlling for the patient’s demographic region, 2014a,b; Tandon et al., 2014]. parental health insurance plan type and parental work industry, KAGELEIRY ET AL. 3 FIG. 1. In this example, a person with Down syndrome was enrolled in an insurance plan during two separate continuous periods. Only the first enrollment period was used for our calculations; data were included in the 0 to <1, 1 to <3, and 3 to <5 age segments. [Color figure can be viewed at wileyonlinelibrary.com]. patient-age observations in the DS, and potential control group All analyses were performed using SAS Version 9.3 (SAS Insti- wererequiredtomatchtowithinaquarterofastandarddeviation tute, Cary, NC), and statistical significance was evaluated at the of the estimated propensity score. 0.05 level (two-sided). Outcomes RESULTS Followingthematchingprocess,averageyearlyhealthcareutilization Patient Selection and Characteristics costswerecalculatedforeachpatientwithineachagegroup.Usingall medical and pharmacy claims during the relevant age ranges, costs A total of 5,167 individuals with DS and 5,673,804 individuals included all out-of-pocket co-pay and co-insurance payments to without chromosomal conditions were initially selected after all measure the direct, real-world costs incurred by patients and their inclusion/exclusion criteria were applied (Table I). A majority of parents. All costs were adjusted to 2013 USD using the medical care patients in each initial group had both an identifiable mother and componentoftheConsumerPriceIndex(CPI);costswereannualized father. Patients in the study cohort and potential control groups over the entire relevant age period by summing all incurred costs (by differedonanumberofdimensions(TableII).Agreaternumberof category) and dividing by the length of the individual’s continuous identifiedpatientswithDSweremale(54.1%vs.51.1%,P<0.001); eligibility within the age group [United States Bureau of Labor these individuals also had significantly longer continuous follow- Statistics, 2014]. A similar approach was used to assess the mean up time in their selected observation period (4.05 vs. 2.83 years, annualcoststoprivatethird-partypayers(insurers),bymeasuringall P<0.001). Patients in the study cohort also had parents with a of such incurred payments for the same medical care received by the significantlygreateraverageageatthetimeoftheirbirthcompared selectedpatients.Forthisanalysis,thereal-worldcostsreimbursedby to the initial control group (33.53 vs. 30.12 years, P<0.001). insurers were summarized; these values will differ from the initial Patients in the two initial groups also differed on a number of amounts charged by healthcare providers. otherdemographicfactors,suchasregion,insuranceplantype,and other parental characteristics. Statistical Analyses Followingthematchingprocess,patientswerecomparedwithin patient-age groups on a similar set of demographic characteristics. Prior to matching, descriptive characteristics were compared be- These patient-age cohorts were found to be statistically similar on tween study and potential control individuals. Comparisons most dimensions, where most remaining statistical differences employed Wilcoxon rank-sum tests for continuous variables were not clinically meaningful (see Supplementary Materials). andchi-squaretestsforcategoricalvariables.Followingthematch- ing process, individuals with DS and their matched controls were Out-of-Pocket Costs for Patients With DS compared within each observed age category on a subset of the same descriptive characteristics. These comparisons utilized Across all age categories, mean total out-of-pocket annual costs Wilcoxon signed-rank tests for continuous variables and McNe- were found to be greater among individuals with DS compared to mar tests for categorical variables. their matched controls. In terms of incremental costs, annual Average yearly health care utilization costs (both to patients out-of-pocket payments incurred by the cohort with DS relative and third-party payers) were compared between the respective tocontrolswithoutDSwerefoundtobe$1,907higherperyearfor patient-age cohorts with DS and matched control individuals. children <1 year old; $1,673 higher per year for children 1 to <3 Cost categories included inpatient, outpatient, emergency room, yearsold;$1,263higherperyearforchildren3to<5yearsold;$973 home health agency, and other medical costs (e.g., laboratory/ higherperyearforchildren5to<13yearsold;and$537higherper pathology costs). Pharmacy costs were also summarized sepa- yearforpatients13to<18yearsold(allP<0.001,Fig.2).Patients rately. All cost dimensions were compared using Wilcoxon with DS were also found to have significantly higher mean annual signed-rank tests. out-of-pocket costs than their matched controls within individual 4 AMERICAN JOURNAL OF MEDICAL GENETICS PART A TABLE I. Selection of Individuals With DS and Individuals Without Chromosomal Anomalies Selection criteria Count Selection of individuals with DS (pre-match study group) Step 0. All beneficiaries 18,028,545 Step 1. Identify individuals with at least one diagnosis for DSa,b 7,918 Step 2. Identify children with co-enrolled parent a) Individual is enrolled as a childc 5,230 b) Individual has an identifiable mother and/or fatherd,e 5,167 Mother: Father: 4,879 4,621 Selection of individuals without DS (pre-match potential control group) Step 0. All beneficiaries 18,028,545 Step 1. Identify individuals without any diagnoses for chromosomal anomaliesa,f 18,008,380 Step 2. Identify children with co-enrolled parent a) Individual is enrolled as a childc 5,739,023 b) Individual has an identifiable mother and/or fatherd,e 5,673,804 Mother: Father: 5,230,957 4,998,205 DS, Down syndrome. a Diagnoses assessed in medical claims from Q1 1999 to Q1 2013. b DS was defined as ICD-9-CM: 758.0x. cIndividual was classified as a child of a beneficiary on their insurance enrollment. d The oldest woman/man classified as a subscriber or spouse on a family insurance enrollment is assumed to be the mother/father (respectively). e Individuals are also required to have identifiable information used in matching, including gender, date of birth, region, insurance plan type, and parent work industry. fChromosomal anomalies were defined as ICD-9-CM: 758.xx. cost categories for all age groups. The greatest incremental costs for patients with DS between birth and age 18 can be calculated were found to be inpatient costs in the 1st year of life ($1,183 vs. assuming that incremental costs within each age category are $259, incremental: $925, P<0.001) and outpatient costs in later constant for each year within that time period. This approach years (incremental costs ranging $183–$623, all P<0.001) suggeststhatindividualswithDSareobservedtoincurincremental (Table III). Incremental out-of-pocket pharmacy expenditures total healthcare expenditures of $230,043 between birth and age and emergency room costs were smaller and similar across all 18 years when compared to similar individuals without chromo- age groups for patients with DS. somal conditions. When the costs are amortized over 18 years, ThetotalincrementalcostincurredtopatientswithDSbetween third-party companies pay, on average, an additional $1,065 per birth and age 18 can be calculated assuming that incremental costs month for medical expenses for persons with DS. withineachagecategoryareconstantforeachyearwithinthattime period. This approach suggests that individuals with DS are DISCUSSION observed to incur incremental out-of-pocket medical costs of $18,248 between birth and age 18 years when compared to similar Caregivers who have children with DS paid, on average, an addi- individuals without chromosomal conditions. When the costs are tional$18,248inout-of-pocketmedicalexpenseswhencomparedto amortizedover18years,parentswhohavechildrenwithDSpay,on parentsofchildrenwithoutDSbetweentheagesofbirthto18years. average, an additional $84 per month for medical expenses. Whenamortizedoverthistimeperiod,thisamountstoanextra$84 per month. These medical expenses included inpatient costs, out- Total Healthcare Expenditures to Third-Party patientcosts,emergencyroomvisits,homehealthagencycosts,and Payers outpatient pharmacy costs. The most costly services were inpatient costs in the 1st year of life when the need for surgery is greatest. Across all age categories, mean total healthcare expenditures were During this time, approximately 40–50% of infants with DS are found to be significantly greater among individuals with DS identified as having a cardiac condition, many of whom will need comparedtotheirmatchedcontrols.Theannualincrementalcosts surgicalrepair[Bull,2011].Gastrointestinalcomplicationssuchasa associated withDSincurredtothird-partypayerswerefoundtobe tracheoesophagealfistula,duodenalobstruction,Hirschsprungdis- $80,864 per year for children <1 year old; $18,950 per year for ease, and an imperforate anus may also necessitate surgery children 1 to <3 years old; $11,228 per year for children, 3 to <5 [Freeman et al., 2009]. Likewise, cataracts can also occur during years old; $7,586 per year for children, 5 to <13 years old; and thistimeframe,requiringsurgicalcorrectionbyanophthalmologist $5,627 per year for patients, 13 to <18 years old (all P<0.001, see [Creavin and Brown, 2009]. Still others develop infantile spasms, TableIV).Thetotalincrementalcostincurredtothird-partypayers requiring hospitalization for medication control of their seizures
no reviews yet
Please Login to review.