Advertisement

Productivity Loss and Indirect Burden of Cyclic Vomiting Syndrome in the United States

  • Author Footnotes
    ∗ Dr Song was employed at IBM Watson Health at the time this study was executed. She is currently employed at Regeneron, Tarrytown, NY, USA.
    Xue Song
    Footnotes
    ∗ Dr Song was employed at IBM Watson Health at the time this study was executed. She is currently employed at Regeneron, Tarrytown, NY, USA.
    Affiliations
    IBM Watson Health, Cambridge, Massachusetts
    Search for articles by this author
  • Yaozhu J. Chen
    Affiliations
    Takeda Development Center Americas, Inc., Cambridge, Massachusetts
    Search for articles by this author
  • Allison Perry
    Affiliations
    IBM Watson Health, Cambridge, Massachusetts
    Search for articles by this author
  • Jerry Kagan
    Affiliations
    IBM Watson Health, Cambridge, Massachusetts
    Search for articles by this author
  • Sanjay Bhandari
    Affiliations
    Division of Gastroenterology and Hepatology, Medical College of Wisconsin, Milwaukee, Wisconsin
    Search for articles by this author
  • Cristina Almansa
    Affiliations
    Takeda Development Center Americas, Inc., Cambridge, Massachusetts
    Search for articles by this author
  • Camilla Richmond
    Affiliations
    Takeda Development Center Americas, Inc., Cambridge, Massachusetts
    Search for articles by this author
  • David J. Levinthal
    Affiliations
    Division of Gastroenterology and Hepatology and Nutrition, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
    Search for articles by this author
  • Author Footnotes
    ‡ Dr Venkatesan was employed at the Medical College of Wisconsin at the time this study was executed. She is currently employed at the Ohio State Wexner Medical Center, Columbus, OH, USA.
    Thangam Venkatesan
    Correspondence
    Correspondence: Address correspondence to: Thangam Venkatesan, MD, The Ohio State University College of Medicine, 395 W 12th Avenue, Columbus, Ohio 43210.
    Footnotes
    ‡ Dr Venkatesan was employed at the Medical College of Wisconsin at the time this study was executed. She is currently employed at the Ohio State Wexner Medical Center, Columbus, OH, USA.
    Affiliations
    Division of Gastroenterology and Hepatology, Medical College of Wisconsin, Milwaukee, Wisconsin
    Search for articles by this author
  • Author Footnotes
    ∗ Dr Song was employed at IBM Watson Health at the time this study was executed. She is currently employed at Regeneron, Tarrytown, NY, USA.
    ‡ Dr Venkatesan was employed at the Medical College of Wisconsin at the time this study was executed. She is currently employed at the Ohio State Wexner Medical Center, Columbus, OH, USA.
Open AccessPublished:July 20, 2022DOI:https://doi.org/10.1016/j.gastha.2022.06.017

      Background and Aims

      To quantify the indirect burden of cyclic vomiting syndrome (CVS), we assessed work-related productivity loss in patients with CVS and caregivers using large-sized databases in the United States.

      Methods

      Patients aged 18–64 years with full-time employment in MarketScan Commercial and Health and Productivity Management Databases were selected if they had ≥1 inpatient or ≥2 outpatient claims for CVS between 2008 and 2018 and continuous enrollment of ≥6 months before and ≥3 months after the initial CVS diagnosis. CVS caregivers were adults with full-time employment and also having dependent(s) with CVS. Propensity scores via multivariable regressions were used to match patients with CVS and their caregivers to non-CVS controls. Productivity loss was assessed by short-term disability (STD) and absenteeism (ABS) days, and the associated costs were also calculated. Differences between the matched cohorts were regarded as the burden attributable to CVS.

      Results

      Patients with CVS had longer annualized STD (21.1 vs 7.0, P < .001) and ABS days (26.4 vs 22.8, P < .05) than their matched controls. CVS caregivers had more annualized STD (3.9 vs 2.6, P < .001) and ABS days (20.9 vs 19.5, P < .05) than controls. Productivity loss costs for STD or ABS days were greater for patients with CVS and caregivers. Annualized health-care resource utilization (inpatient, emergency room, outpatient) was 5.2–6.0 times higher in patients with CVS (P < .001).

      Conclusion

      CVS is associated with higher productivity loss due to STD/ABS and, therefore, greater indirect costs for patients and caregivers. Further research is needed to assess the full societal burden of CVS. More effective interventions may reduce the disease burden.

      Keywords

      Abbreviations used in this paper:

      ABS (absenteeism), CVS (cyclic vomiting syndrome), DCI (Deyo-Charlson Comorbidity Index), ER (emergency room), FTE (full-time employee), HRU (health-care resource utilization), HPM (Health Productivity Management), ICD-9 (International Classification of Diseases-Ninth Revision), ICD-10 (International Classification of Diseases-Tenth Revision), STD (short-term disability)

      Introduction

      Cyclic vomiting syndrome (CVS) is a chronic disorder of gut-brain interaction, characterized by episodic nausea and repetitive vomiting.
      • Stanghellini V.
      • Chan F.K.
      • Hasler W.L.
      • et al.
      Gastroduodenal disorders.
      It is estimated to affect ∼2% of children, with a similar prevalence in adults.
      • Sagar R.C.
      • Sood R.
      • Gracie D.J.
      • et al.
      Cyclic vomiting syndrome is a prevalent and under-recognized condition in the gastroenterology outpatient clinic.
      • Sperber A.D.
      • Bangdiwala S.I.
      • Drossman D.A.
      • et al.
      Worldwide prevalence and burden of functional gastrointestinal disorders, results of Rome Foundation Global Study.
      • Aziz I.
      • Palsson O.S.
      • Tornblom H.
      • et al.
      Epidemiology, Clinical Characteristics, and Associations for Symptom-Based Rome IV Functional Dyspepsia in Adults in the USA, Canada, and the UK: a Cross-Sectional Population-Based Study.
      The pathogenesis of CVS is unknown and appears to be multifactorial, with several potential disease mechanisms.
      • Sunku B.
      Cyclic vomiting syndrome: a disorder of all ages.
      ,
      • Hasler W.L.
      • Levinthal D.J.
      • Tarbell S.E.
      • et al.
      Cyclic vomiting syndrome: pathophysiology, comorbidities, and future research directions.
      This condition is further complicated by comorbid conditions including anxiety, depression, autonomic dysfunction, migraine, and the need for management with lifestyle interventions, supportive care, and abortive and/or prophylactic medications.
      • Li B.U.K.
      Managing cyclic vomiting syndrome in children: beyond the guidelines.
      ,
      • Venkatesan T.
      • Levinthal D.J.
      • Tarbell S.E.
      • et al.
      Guidelines on management of cyclic vomiting syndrome in adults by the American Neurogastroenterology and Motility Society and the Cyclic Vomiting Syndrome Association.
      Several published studies in patients with CVS have noted functional disability, decreased quality of life, and increased health-care costs.
      • Venkatesan T.
      • Levinthal D.J.
      • Tarbell S.E.
      • et al.
      Guidelines on management of cyclic vomiting syndrome in adults by the American Neurogastroenterology and Motility Society and the Cyclic Vomiting Syndrome Association.
      • Lee L.Y.
      • Abbott L.
      • Mahlangu B.
      • et al.
      The management of cyclic vomiting syndrome: a systematic review.
      • Taranukha T.
      • Charan Suresh Kumar V.
      • Seamon A.
      • et al.
      Depression, young age, chronic marijuana use, and interepisodic symptoms predict psychological distress in patients with cyclic vomiting syndrome.
      • Tarbell S.E.
      • Li B.U.
      Health-related quality of life in children and adolescents with cyclic vomiting syndrome: a comparison with published data on youth with irritable bowel syndrome and organic gastrointestinal disorders.
      • Tarbell S.E.
      • Li B.U.
      Anxiety Measures predict health-related quality of life in children and adolescents with cyclic vomiting syndrome.
      • Tarbell S.E.
      • Millar A.
      • Laudenslager M.
      • et al.
      Anxiety and physiological responses to the trier social stress test for children in adolescents with cyclic vomiting syndrome.
      • Ye Z.
      • Xue A.
      • Huang Y.
      • et al.
      Children with cyclic vomiting syndrome: phenotypes, disease burden and mitochondrial DNA analysis.
      • Levinthal D.J.
      • Romutis S.
      • Rajalaban A.
      • et al.
      Greater intolerance to uncertainty predicts poorer quality of life in adults with cyclic vomiting syndrome.
      CVS causes substantial morbidity, with half of the patients requiring emergency room (ER) care and intravenous therapy or inpatient admissions, leading to significant time lost from work or school.
      • Fitzpatrick E.
      • Bourke B.
      • Drumm B.
      • et al.
      Outcome for children with cyclical vomiting syndrome.
      ,
      • Li B.U.
      • Balint J.P.
      Cyclic vomiting syndrome: evolution in our understanding of a brain-gut disorder.
      Such episodic incapacity in adult patients with CVS has led to frequent work absences, delays in education, job loss, and disability.
      • Levinthal D.J.
      • Romutis S.
      • Rajalaban A.
      • et al.
      Greater intolerance to uncertainty predicts poorer quality of life in adults with cyclic vomiting syndrome.
      ,
      • Li B.U.
      • Balint J.P.
      Cyclic vomiting syndrome: evolution in our understanding of a brain-gut disorder.
      Studies on CVS in children estimate an average loss of school days per year between 20 and 24 due to ER visits and scheduling of repetitive testing.
      • Fitzpatrick E.
      • Bourke B.
      • Drumm B.
      • et al.
      Outcome for children with cyclical vomiting syndrome.
      • Li B.U.
      • Balint J.P.
      Cyclic vomiting syndrome: evolution in our understanding of a brain-gut disorder.
      • Foreman M.S.
      • Camp T.
      Cyclic vomiting syndrome.
      Because of these repeated school absences, many children have required home tutoring or home schooling, which places additional burdens on their caregivers and is negatively associated with parent emotional functioning.
      • Wang-Hall J.
      • Li B.U.K.
      • Tarbell S.E.
      Family health-related quality of life in Pediatric cyclic vomiting syndrome.
      One study found that anxiety and missing school days in children living with CVS strongly predicted lower family health-related quality of life, affecting “family physical functioning, family communication, and family daily activities.”
      • Wang-Hall J.
      • Li B.U.K.
      • Tarbell S.E.
      Family health-related quality of life in Pediatric cyclic vomiting syndrome.
      Despite the recognition that patients with CVS and caregivers experience substantial work-related productivity loss and associated costs, the extent of this impact has not been well assessed and quantified. To evaluate the magnitude of indirect burden of CVS, this study assessed work-related productivity loss in adults with CVS and their caregivers using large-scaled claims and productivity databases in the United States.

      Methods

      Data Sources

      This observational retrospective cohort analysis utilized deidentified US administrative claims data covering from July 1, 2007, to December 31, 2018, (study window) in 2 IBM MarketScan Databases: Commercial Claims and Encounters (Commercial) Database and the Health Productivity and Management (HPM) Database. The Commercial Database contains data on the inpatient, outpatient, and outpatient prescription drug use of employees and their dependents, covered under a variety of fee-for-service and managed care health plans. The HPM Database contains lost work days, short-term disability (STD), long-term disability, and workers’ compensation data from a subset of MarketScan employer clients. Data elements for a same employee can be fully linked in both databases.
      All study data were obtained and classified by International Classification of Diseases, 9th and 10th Revision, Clinical Modification (ICD-9-CM and ICD-10-CM) codes, Current Procedural Terminology 4th edition codes, Healthcare Common Procedure Coding System codes, and National Drug Codes.

      Patient Selection: CVS Patients vs non-CVS Controls

      Data of patients with ≥1 inpatient or ≥2 outpatient claims, on different dates, with a diagnosis for CVS in any position (ICD-9-CM: 536.2; ICD-10-CM: G43.A0, G43.A1) between January 1, 2008, and December 31, 2018, (patient selection window) were extracted. Index date was the date of first CVS claim in the patient selection window. We included patients who were aged 18–64 years on index date, as active full-time employees (FTEs) (ie, not dependents), and having continuous enrollment with medical and prescription coverages for ≥6 months before the index date (preindex or baseline period) and ≥3 months after the index date (postindex or follow-up period). To ensure that selected patients were newly diagnosed with CVS during the study period, those with any diagnosis claim of CVS or vomiting during the preindex period were excluded. Patients with claims for pregnancy or birth delivery during the study window were also excluded. Two subcohorts of patients with CVS were created: one consisting of patients with work absenteeism (ABS) eligibility, and the other with STD eligibility. We included patients who had ≥6-month preindex and ≥3-month postindex ABS or STD eligibility in each subcohort, respectively (Figure 1). Both subcohorts were followed up for variable lengths of time, until the end of their eligibility with employer-based work absence or STD benefits, disenrollment in the databases, or the end of study window (December 31, 2018), whichever is earlier.
      Figure thumbnail gr1
      Figure 1Cyclic vomiting syndrome (CVS) patient attrition. 1For each patient with CVS, the number of days between the index date and January 1, 2008, was calculated and referred to as the “interval pool.” For each control patient, a number was randomly drawn from that interval pool, and the index date equals that number plus January 1, 2008.
      A non-CVS cohort was selected from a 10% random sample of patients without CVS in the linked Commercial and HPM Databases, with ≥1 year of benefit eligibility for absence or STD between January 1, 2008, and December 31, 2018. Index date was randomly assigned to match the distribution of index dates in the CVS cohorts. Controls were required to meet all inclusion and exclusion criteria of patients with CVS.
      Propensity score matching was implemented for each CVS patient to up to 3 corresponding non-CVS controls, based on their baseline demographics, Deyo-Charlson Comorbidity Index,
      • Deyo R.A.
      • Cherkin D.C.
      • Ciol M.A.
      Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases.
      clinical characteristics, and baseline work absence or STD days. The balances of postmatching cohorts were evaluated using standardized mean differences, with a threshold of its absolute value <10%, set a priori, to indicate balance between cohorts.

      Patient Selection: CVS Caregivers vs non-CVS Controls

      Caregivers of patients with CVS were also identified: aged 18–64 years, with FTE and benefit eligibility for absence or STD, and having a dependent (ie, child or spouse) who had ≥1 inpatient or ≥2 nondiagnostic outpatient claims for CVS in the selection window but no claim in the preindex period. Index date was the date of first documented CVS diagnosis. We included caregivers who had ≥6 months of preindex and ≥3 months of postindex continuous enrollment of medical and pharmacy benefits. Caregivers were excluded from analysis if they had any claims for CVS, pregnancy, or delivery during the study period. Caregiver controls are those who had no CVS diagnosis, pregnancy or delivery, and no family members with CVS (Figure A1). CVS caregivers and their non-CVS controls were matched by propensity scores from multivariable regressions, which control for baseline characteristics including demographics, number of family members, and baseline lost workdays. Each CVS caregiver was matched with up to 3 controls, with the same standardized mean difference threshold to assess balance level.

      Baseline Characteristics

      For patients with CVS and controls, we assessed their baseline demographic characteristics, including age, sex, geographic region, population density, and insurance plan type on the index date. Duration of follow-up was also captured. We measured baseline clinical characteristics during the 6-month preindex period and included comorbidities comprising the Deyo-Charlson Comorbidity Index, as well as conditions identified in the literature as having a higher burden in patients with CVS: abdominal pain, anxiety (including panic disorder), autonomic dysfunction, cannabis abuse/use, cardiac conditions and risks, depression, fibromyalgia, gastroesophageal reflux disease, gastroparesis, irritable bowel syndrome, migraine, nausea, and seizure. For the CVS caregiver analysis, demographic characteristics, including age, sex, geographic region, population density, and insurance plan type were assessed on the index date. Number of family members was also recorded.

      Baseline Productivity Loss

      We recorded the number of lost workdays or STD days for patients with CVS, their caregivers, and their corresponding controls during the preindex period.

      Productivity Loss and Indirect Costs in the Follow-up Period

      Number and proportion of patients with an absence or STD claim were reported during the follow-up period. We annualized number of lost workdays and associated costs (per-patient per-year) to account for variable length of follow-up period. Indirect costs associated with ABS were calculated as number of absent days multiplied by a wage constant of $177.2 per day (ie, $22.15 per hour), which is an equivalent to the median hourly rate for the employed full-time US workers in 2018.
      Median weekly earnings of full-time wage and salary workers by detailed occupation and sex. U.S. Bureau of Labor Statistics.
      Indirect costs associated with STD were calculated using the same wage constant but multiplied by a discounting factor of 60%, which is the typical percentage of an employee’s income that STD policies compensate for.

      All-Cause Health-Care Resource Utilization in the Follow-up Period for Patients With CVS and Controls

      Among the patients with CVS and their matched controls included in this study, we also compared their all-cause health-care resource utilization (HRU), including ER visits, inpatient admissions, outpatient services (such as physician office visits and other outpatient visits), and outpatient pharmacy claims. The proportion of patients with any claim as well as the number of annualized visits or services in each care setting were captured; in addition, length of stay in days per admission was also reported for the inpatient setting. These assessments on HRU provide additional data to inform to what extent the impact of CVS on ABS/STD may relate to the elevated use of medical services.

      Statistical Analysis

      For all baseline characteristics, frequencies and percentages were reported for categorical variables, and mean and standard deviation were reported for continuous variables. The alpha level for all statistical tests was 0.05. All analyses were conducted using WPS version 4.1 (World Programming, United Kingdom). Data are owned by IBM Watson Health and can be accessed through a licensing agreement https://www.ibm.com/products/marketscan-research-databases.

      Results

      Patients With CVS vs non-CVS: Demographics and Clinical Characteristics

      A total of 7080 patients with CVS were identified (Figure 1). After applying study selection criteria and propensity score matching process, 2502 CVS cases and 7127 controls were included in the STD analysis, and 348 CVS cases and 1014 controls were included in the ABS analysis (Figure 1).
      A majority of the patients in CVS case and matched non-CVS control cohorts were female (56.6%–56.9% female), while more men were seen in the final ABS cohorts (42.8%–45.3% female) (Table 1). The STD and ABS cohorts were fairly similar in age (mean 43.7–46.0 years) and prevalence of other digestive problems, such as abdominal pain (29.5%–38.1%) and gastroesophageal reflux disease (12.9%–17.6%). Also, prevalent comorbidities among the patients in CVS cohorts and their matched non-CVS controls include cardiac conditions/risks (36.1%–36.8%), anxiety (11.2%–13.0%), and depression (10.6%–12.1%). In the STD analysis, close to half of the cases and controls were from the South region (44.4%–45.4%); while in the ABS analysis, more patients were from South or West regions (33.8%–34.8% South, 33.6%–35.0% West). In addition, the lengths of follow-up in databases were on average 2.4 years for STD and 3.1 years for ABS.
      Table 1Matched CVS Patient Demographics,
      Demographics were captured on index.
      Baseline Clinical Characteristics, and Productivity Loss
      Baseline clinical characteristics and productivity loss were captured during the 6-month preindex period.
      CharacteristicsShort-term disability (STD)Absenteeism (ABS)
      CVS patientsNon-CVS controls% Standardized differenceCVS patientsNon-CVS controls% Standardized difference
      N = 2502N = 7127N = 348N = 1014
      Age (mean, SD)43.7 ± 10.944.9 ± 10.910.6946.0 ± 11.245.7 ± 11.12.23
      Age group (N, %)
       18–30370 (14.8)941 (13.2)4.5743 (12.4)123 (12.1)0.69
       31–44905 (36.2)2415 (33.9)4.79104 (29.9)309 (30.5)1.28
       45–54729 (29.1)2151 (30.2)2.29106 (30.5)322 (31.8)2.80
       55–64498 (19.9)1620 (22.7)6.9195 (27.3)260 (25.6)3.76
      Sex (N, %)
       Male1086 (43.4)3069 (43.1)0.7199 (57.2)555 (54.7)4.9
       Female1416 (56.6)4058 (56.9)0.7149 (42.8)459 (45.3)4.9
      Days of follow-up (mean, SD)
      Length of follow-up comprises the time from index until the end of follow-up due to end of eligibility, enrollment, or study period (December 31, 2018).
      874.9 (729.5)876.3 (669.0)0.201093.3 (886.9)1179.8 (837.1)10.03
       Median663.5 (0.0)687.0 (0.0)846.5 (0.0)1058.5 (0.0)
      DCI (mean, SD)1.0 (2.0)0.9 (1.8)5.461.1 (2.1)0.9 (1.9)8.70
      Baseline conditions (N, %)
      Comorbid conditions identified in the literature as having a high burden in CVS patients.6,10,11
      ,
      Additional baseline comorbidities captured include autonomic dysfunction, cannabis abuse/use, and seizure, which occurred at a rate of less than 2.0%.
       Abdominal pain954 (38.1)2570 (36.1)4.29109 (31.3)299 (29.5)3.99
       Anxiety314 (12.6)926 (13.0)1.3339 (11.2)115 (11.3)0.42
       Panic disorder37 (1.5)97 (1.4)1.005 (1.4)7 (0.7)7.28
       Cardiac conditions and risks
      Defined as acute myocarditis, acute pericarditis, arrhythmias (including atrial fibrillation and flutter), cardiac arrest, cardiomyopathy, cerebrovascular disease, chronic rheumatic heart disease, conduction disorders, diseases of arteries, arterioles and capillaries, diseases of endocardium, diseases of veins, lymphatic vessels and lymph nodes, heart failure, hypertension, hypotension, ischemic heart disease, paroxysmal tachycardia, and pulmonary heart diseases.
      902 (36.1)2585 (36.3)0.46128 (36.8)366 (36.1)1.43
       Depression289 (11.6)862 (12.1)1.6937 (10.6)117 (11.5)2.89
       Fibromyalgia78 (3.1)217 (3.0)0.429 (2.6)25 (2.5)0.77
       GERD441 (17.6)1248 (17.5)0.3045 (12.9)136 (13.4)1.42
       Gastroparesis44 (1.8)40 (0.6)11.203 (0.9)4 (0.4)5.92
       Irritable bowel syndrome (IBS)49 (2.0)128 (1.8)1.204 (1.1)15 (1.5)2.90
       Migraine174 (7.0)476 (6.7)1.0924 (6.9)76 (7.5)2.32
       Nausea220 (8.8)476 (6.7)7.9222 (6.3)49 (4.8)6.49
      Geographic region (N, %)
       Northeast286 (11.4)895 (12.6)3.539 (11.2)123 (12.1)2.9
       North Central605 (24.2)1669 (23.4)1.870 (20.1)190 (18.7)3.5
       South1137 (45.4)3167 (44.4)2.0121 (34.8)343 (33.8)2.0
       West473 (18.9)1383 (19.4)1.3117 (33.6)355 (35.0)2.9
       Unknown1 (0.0)13 (0.2)4.31 (0.3)3 (0.3)0.2
      Baseline productivity loss
      Baseline clinical characteristics and productivity loss were captured during the 6-month preindex period.
       Days with STD/ABS (mean, SD)7.3 (21.0)6.0 (20.3)6.615.2 (14.5)15.0 (14.2)1.6
      DCI, Deyo-Charlson Comorbidity Index; GERD, gastroesophageal reflux disease.
      a Demographics were captured on index.
      b Baseline clinical characteristics and productivity loss were captured during the 6-month preindex period.
      c Length of follow-up comprises the time from index until the end of follow-up due to end of eligibility, enrollment, or study period (December 31, 2018).
      d Comorbid conditions identified in the literature as having a high burden in CVS patients.
      • Hasler W.L.
      • Levinthal D.J.
      • Tarbell S.E.
      • et al.
      Cyclic vomiting syndrome: pathophysiology, comorbidities, and future research directions.
      ,
      • Taranukha T.
      • Charan Suresh Kumar V.
      • Seamon A.
      • et al.
      Depression, young age, chronic marijuana use, and interepisodic symptoms predict psychological distress in patients with cyclic vomiting syndrome.
      ,
      • Tarbell S.E.
      • Li B.U.
      Health-related quality of life in children and adolescents with cyclic vomiting syndrome: a comparison with published data on youth with irritable bowel syndrome and organic gastrointestinal disorders.
      e Additional baseline comorbidities captured include autonomic dysfunction, cannabis abuse/use, and seizure, which occurred at a rate of less than 2.0%.
      f Defined as acute myocarditis, acute pericarditis, arrhythmias (including atrial fibrillation and flutter), cardiac arrest, cardiomyopathy, cerebrovascular disease, chronic rheumatic heart disease, conduction disorders, diseases of arteries, arterioles and capillaries, diseases of endocardium, diseases of veins, lymphatic vessels and lymph nodes, heart failure, hypertension, hypotension, ischemic heart disease, paroxysmal tachycardia, and pulmonary heart diseases.

      Patients With CVS and non-CVS: Annualized Productivity Loss and Related Costs

      During variable-length follow-up period, a significantly greater proportion of patients with CVS used STD time during follow-up than controls (41.5% vs 18.2%, P < .001), whereas a comparable proportion of cases and controls used ABS time (73.0% vs 76.6%, P = .172). Patients with CVS used 3 times greater average number of annualized STD days (21.1 vs 7.0) and incurred significantly greater corresponding indirect costs ($2245.06 vs $745.60) than controls (P < .001) (Figure 2). Patients with CVS also had a significantly greater number of annualized ABS days (25.9 vs 22.8) and associated costs ($4597.49 vs $4036.43) than controls (P < .05).
      Figure thumbnail gr2
      Figure 2Patient annualized number of short-term disability (STD) and absenteeism (ABS) days and related costs during follow-up. Costs for STD = (number of days lost) × (2018 daily wage of $106.3 [which is 60% of the 2018 US national median daily wage $177.2] for STD). Costs for ABS = (number of days lost) × (2018 US national median daily wage of $177.2 for ABS); ∗P < .001.

      Patients With CVS vs non-CVS: Annualized All-Cause HRU

      Table 2 presents annualized all-cause HRU during the follow-up period for matched patients with CVS and non-CVS controls in the STD and ABS subcohorts. Across all service categories, HRU was higher for patients with CVS than for non-CVS controls (Table 2). Specifically, the proportion of patients with CVS with an inpatient admission was between 5.2 and 6.0 times higher than controls, with a prevalence of 70.5% in cases and 11.7% in controls in the STD subcohort and 68.4% vs 13.0% in the ABS subcohort (P < .001). The average number of postindex inpatient admissions among patients with CVS was 12 times higher than that in controls although mean lengths of stay per admission were comparable. This trend held when comparing mean annual ER visits, outpatient services, and prescription fills (all P < .001).
      Table 2Annualized All-Cause Health-care Utilization in Follow-Up for Matched CVS and Non-CVS Patients
      CharacteristicsShort-term disability (STD)Absenteeism (ABS)
      CVS patientsNon-CVS controlsP valueCVS patientsNon-CVS controlsP value
      N = 2502N = 7127N = 348N = 1014
      Inpatient admissions
       Patients with an admission (N, %)1764 (70.5)836 (11.7)<.001238 (68.4)132 (13.0)<.001
      Avg. length of stay per admission, d (mean ± SD)3.93 ± 3.933.86 ± 4.45.484.08 ± 3.583.75 ± 3.64.14
       Number of admissions, annualized (mean ± SD)1.20 ± 1.980.10 ± 0.54<.0011.17 ± 2.170.09 ± 0.44<.001
      Emergency room (ER) visits
       Patients with an ER visit (N, %)1762 (70.4)2503 (35.1)<.001243 (69.8)392 (38.7)<.001
       Number of ER visits, annualized (mean ± SD)1.92 ± 3.480.40 ± 1.17<.0011.34 ± 2.160.40 ± 1.30<.001
      Outpatient services
       Outpatient office visits
      Patients with an office visit (N, %)2443 (97.6)6658 (93.4)<.001345 (99.1)957 (94.4)<.001
      Number of office visits, annualized (mean ± SD)287.13 ± 347.35187.41 ± 244.11<.001341.10 ± 407.45224.32 ± 243.43<.001
       Other outpatient services
      Patients with other outpatient services (N, %)2476 (99.0)6636 (93.1)<.001346 (99.4)950 (93.7)<.001
      Number of other outpatient services, annualized (mean ± SD)84.82 ± 115.7437.11 ± 55.80<.00184.52 ± 128.5337.31 ± 57.53<.001
      Outpatient pharmacy
       Patients with an outpatient prescription (N, %)2455 (98.1)6602 (92.6)<.001338 (97.1)949 (93.6).01
       Number of outpatient prescriptions, annualized (mean ± SD)33.81 ± 30.1720.66 ± 21.80<.00133.73 ± 28.9622.25 ± 24.88<.001

      Caregiver CVS vs non-CVS: Demographics and Clinical Characteristics

      A total of 92,509 adult family members of patients with CVS were identified (Figure A1). After the employment of selection criteria and propensity score matching, 7342 CVS caregivers and 22,013 non-CVS caregiver controls were included for the STD analysis, and 1318 CVS caregivers and 3942 non-CVS caregiver controls were captured for the ABS analysis.
      The mean ages of CVS caregivers and their matched non-CVS controls were between 43.9 and 46.1 years, and those of their dependents with CVS were between 23.5 and 25.8 years on average (Table 3). Most of the CVS caregivers and non-CVS controls included in this study were male (72.2%–82.9%), with a mean of 2.6–2.7 family members per caregiver. Caregivers and controls were followed up for an average of 2.8–4.0 years.
      Table 3Matched Caregiver Demographics
      Demographics were captured on index.
      and Productivity Loss
      CharacteristicsShort-term disability (STD)Absenteeism (ABS)
      CaregiversCaregiver controls% Standardized differenceCaregiversCaregiver controls% Standardized difference
      N = 7342N = 22,013N = 1318N = 3942
      Age (mean, SD)43.9 (9.1)44.1 (9.0)2.5445.5 (8.9)46.1 (9.0)5.84
      Age group (N, %)
       18–30566 (7.7)1643 (7.5)0.9380 (6.1)203 (5.1)4.00
       31–443289 (44.8)9856 (44.8)0.05493 (37.4)1416 (35.9)3.08
       45–542458 (33.5)7365 (33.5)0.04520 (39.5)1578 (40.0)1.18
       55–641029 (14.0)3149 (14.3)0.83225 (17.1)745 (18.9)4.76
      Sex (N, %)
       Male5300 (72.2)16,026 (72.8)1.41092 (82.9)3268 (82.9)0.1
      Days of follow-up (mean, SD)
      Length of follow-up comprises the time from index until the end of follow-up due to end of eligibility, enrollment, or study period (December 31, 2018).
      1158.7 (851.3)1037.2 (737.5)NA1462.4 (973.5)1272.1 (848.0)NA
       Median975.0884.01326.01205.5
      No. of family members (mean, SD)2.7 (1.3)2.7 (1.3)0.232.6 (1.3)2.6 (1.3)0.84
       Median3.03.03.03.0
      CVS patient characteristics
      Relationship to caregiver also evaluated employee and dependent relation unknown, which had an N (%) of 0 for both STD and ABS CVS caregiver cohorts.
       Age (mean, SD)23.5 (18.1)------25.8 (18.1)------
       Sex (N, %)
      Male2706 (36.9)------411 (31.2)------
       Relationship to caregiver (N, %)
      Spouse2615 (35.6)------514 (39.0)------
      Child/Other4727 (64.4)------804 (61.0)------
      Baseline productivity loss
      Baseline clinical characteristics and productivity loss were captured during the 6-mo preindex period.
       Days with STD/ABS (mean, SD)1.9 (10.7)1.6 (10.5)2.612.2 (10.5)11.9 (10.7)2.9
      a Demographics were captured on index.
      b Baseline clinical characteristics and productivity loss were captured during the 6-mo preindex period.
      c Length of follow-up comprises the time from index until the end of follow-up due to end of eligibility, enrollment, or study period (December 31, 2018).
      d Relationship to caregiver also evaluated employee and dependent relation unknown, which had an N (%) of 0 for both STD and ABS CVS caregiver cohorts.

      Caregiver CVS vs non-CVS: Annualized Productivity Loss and Related Costs

      During the follow-up, more CVS caregivers used STD time than their controls (16.1% vs 10.3%, P < .001), whereas the proportions with ABS time use were more comparable between CVS caregivers and non-CVS controls (75.2% vs 74.7%, P = .741). In comparison to the matched non-CVS controls, CVS caregivers had a higher number of STD and ABS days (Figure 3), and consequently, their annualized productivity loss in workdays and associated indirect costs were 1.5 times higher for STD (3.9 vs 2.6 days, $410.04 vs $273.60; P < .001) and 1.1 times higher for ABS (20.9 vs 19.5, $3700.20 vs $3460.92; P < .05).
      Figure thumbnail gr3
      Figure 3Caregiver annualized number of short-term disability (STD) and absenteeism (ABS) days and related costs during follow-up. Costs for STD = (number of days lost) × (2018 daily wage of $106.3 [which is 60% of the 2018 US national median daily wage $177.2] for STD). Costs for ABS = (number of days lost) × (2018 US national median daily wage of $177.2 for ABS); ∗P < .01; ∗∗P < .05.

      Conclusions

      While the previous literature established that CVS imposes significant burden due to health-care use and costs, increased work and school ABS, as well as reduced quality of life, these prior studies also had limitations, such as a small sample size, reliance on survey data, or restriction to specific settings of care.
      • Lee L.Y.
      • Abbott L.
      • Mahlangu B.
      • et al.
      The management of cyclic vomiting syndrome: a systematic review.
      ,
      • Taranukha T.
      • Charan Suresh Kumar V.
      • Seamon A.
      • et al.
      Depression, young age, chronic marijuana use, and interepisodic symptoms predict psychological distress in patients with cyclic vomiting syndrome.
      ,
      • Tarbell S.E.
      • Li B.U.
      Anxiety Measures predict health-related quality of life in children and adolescents with cyclic vomiting syndrome.
      ,
      • Tarbell S.E.
      • Millar A.
      • Laudenslager M.
      • et al.
      Anxiety and physiological responses to the trier social stress test for children in adolescents with cyclic vomiting syndrome.
      ,
      • Wang-Hall J.
      • Li B.U.K.
      • Tarbell S.E.
      Family health-related quality of life in Pediatric cyclic vomiting syndrome.
      ,
      • Buono J.L.
      • Carson R.T.
      • Flores N.M.
      Health-related quality of life, work productivity, and indirect costs among patients with irritable bowel syndrome with diarrhea.
      • Lucia-Casadonte C.J.
      • Whaley K.G.
      • Chogle A.S.
      Yield and costs of evaluating children with cyclic vomiting syndrome.
      • Venkatesan T.
      • Tarbell S.
      • Adams K.
      • et al.
      A survey of emergency department use in patients with cyclic vomiting syndrome.
      • Bhandari S.
      • Venkatesan T.
      Clinical characteristics, comorbidities and hospital outcomes in hospitalizations with cyclic vomiting syndrome: a nationwide analysis.
      Our study methods leveraged population-level administrative databases to quantify the indirect burden of work productivity loss and associated costs which are attributable to CVS. The results of the current study clearly demonstrate that CVS is associated with significant productivity loss and a substantial indirect burden in the United States.
      We found that indirect costs are an important component of the economic burden of CVS. In our study, newly diagnosed patients had an annualized average of 14.1 more days than non-CVS controls for STD (resulting in 3 times higher costs, P < .001) and 3.6 days more for ABS (leading to 1.1 times higher costs, P < .05). Increased workday loss in patients with CVS may relate to their higher use of health-care resource, primarily due to ER and inpatient care. We found that patients with CVS were 5.2–6.0 times more likely to have an inpatient admission (70.5% vs 11.7% in STD and 68.4% vs 13.0% in ABS, P < .001), and their average number of annualized inpatient admissions were 12 times higher than those of controls (1.20 vs 0.10 for STD and 1.17 vs 0.09 for ABS, P < .001).
      Productivity loss patterns in caregivers were similar. CVS caregivers used, on average, 1.3 more STD days and were absent for 1.4 more days from work, which translated into an annual average of 1.5 times higher indirect costs in STD ($410 vs $274, P < .001) and 1.1 times higher costs in ABS ($3700 vs $3,461, P < .05). These results in combination suggest a substantial productivity loss due to CVS, which imposes an economic burden not only for patients and their caregivers but to society at large.
      This study has several notable strengths. The data source was based on national databases from many large employers and represents a broad range of insurance plans, which result in large sample sizes that are much larger than those used in the previous CVS literature. As the data are suitable to support the estimations at population level, this study is among the first population-level analyses to quantify the indirect burden of CVS on patients and caregivers. Also, this study has the strength in design to quantify the indirect burden of CVS based on the differences of all-cause ABS and STD between CVS cases and non-CVS controls who were matched via propensity scores based on patients’ baseline characteristics.
      However, several limitations need to be considered when interpreting the results. First, like in any claims-based study, the databases were developed for administrative purposes rather than for research and, thus, are subject to coding misclassifications and data entry errors. Additionally, prior to 2016, there was no ICD diagnosis code specific to CVS. To address this issue, our analysis focused entirely on incident cases of CVS, as the additional restrictions allowed us to more specifically identify CVS patients. As additional time elapses, future analysis should be performed using only post-2016 data in which longstanding CVS patients may be identified. Furthermore, future work should be performed to validate via chart review the ICD coding for CVS in a linked EMR-claims database. The reliance of claims databases on ICD coding makes the data readily accessible as compared to Electronic Medical Records (EMR) data. Therefore, despite its limitations, claims databases provide a strong framework for these future analyses.
      Second, the study focused on full-time employment and did not include other employment status (eg, part time, transition from full time to part time, or in early retirement) or no employment (eg, homemaker) due to data availability. Therefore, this study tends to undercapture employment types associated with productivity loss. Additionally, CVS is known to be more common in female patients, yet in this study, we see a higher proportion of male patients in the ABS analysis. This may be associated with a higher proportion of men in the underlying HPM Database, reflecting the sex distribution of FTEs in the United States.
      • Sunku B.
      Cyclic vomiting syndrome: a disorder of all ages.
      ,
      U.S. Bureau of Labor Statistics
      Full-time / part-time employment. U.S. Department of Labor.
      With regard to the sex distribution in the caregiver burden analysis, women are more likely than men to drop out of the work force to take care of a sick family member,
      • Parker K.
      Women more than men adjust their careers for family life. Pew Research Center.
      which was not captured by the HPM Database, leading to an underestimation of the indirect burden of CVS if family member caregivers were not FTEs at the same time. Furthermore, the approach of propensity score matching may underestimate the attributable productivity loss, as the cohort-matching process excluded the most severe patients with CVS who may not find a well-matched non-CVS control. On the other hand, the matched non-CVS controls tend to be sicker than the general population, which may further narrow the difference between CVS and controls and indicate that the burden estimates from this study are conservative. In addition, it often takes months or even years for a CVS patient to be diagnosed, so this study focusing on the newly diagnosed patients does not reflect the productivity loss prior to the first diagnosis of CVS. Another type of productivity loss, also underestimated in this study due to lack of data, is presenteeism (ie, being at work while ill and performing at a lower level than usual). However, since STD benefits are typically used for one’s own medical condition, rather than caring for a sick family member, the STD days of caregivers may overestimate the impact of CVS on indirect costs of caregivers. Finally, our estimates on CVS indirect burden reflect those in a commercially insured population with full-time employment, which may not be generalizable to those with other or no insurance, whose employers do not provide STD or ABS benefits, part-time employees, or homemakers.
      Patients with CVS and their caregivers experience substantial productivity loss with economic implications that extend well beyond the direct costs for medical services and treatment. This study adds quantifiable evidence on the impact of CVS and highlights the critical need for new and more-effective treatment options and management strategies. These may reduce the burden due to CVS on patients, their caregivers, employers, and society at large.

      Acknowledgments:

      The authors would like to acknowledge David Smith of IBM Watson Health for providing the propensity matching.

      Supplementary Materials

      • Figure A1

        Caregiver Attrition. 1Match eligible controls were captured from eligible non-CVS control patients from the CVS patient analysis, and their adult family members were identified. 2Caregivers were followed up from the index date until the earlier of the end-of-work ABS or STD eligibility, disenrollment in the database, or end of the study period (December 31, 2018).

      References

        • Stanghellini V.
        • Chan F.K.
        • Hasler W.L.
        • et al.
        Gastroduodenal disorders.
        Gastroenterology. 2016; 150: 1380-1392
        • Sagar R.C.
        • Sood R.
        • Gracie D.J.
        • et al.
        Cyclic vomiting syndrome is a prevalent and under-recognized condition in the gastroenterology outpatient clinic.
        Neurogastroenterol Motil. 2018; 30
        • Sperber A.D.
        • Bangdiwala S.I.
        • Drossman D.A.
        • et al.
        Worldwide prevalence and burden of functional gastrointestinal disorders, results of Rome Foundation Global Study.
        Gastroenterology. 2021; 160: 99-114.e3
        • Aziz I.
        • Palsson O.S.
        • Tornblom H.
        • et al.
        Epidemiology, Clinical Characteristics, and Associations for Symptom-Based Rome IV Functional Dyspepsia in Adults in the USA, Canada, and the UK: a Cross-Sectional Population-Based Study.
        Lancet Gastroenterol Hepatol. 2018; 3: 252-262
        • Sunku B.
        Cyclic vomiting syndrome: a disorder of all ages.
        Gastroenterol Hepatol (N Y). 2009; 5: 507-515
        • Hasler W.L.
        • Levinthal D.J.
        • Tarbell S.E.
        • et al.
        Cyclic vomiting syndrome: pathophysiology, comorbidities, and future research directions.
        Neurogastroenterol Motil. 2019; 31: e13607
        • Li B.U.K.
        Managing cyclic vomiting syndrome in children: beyond the guidelines.
        Eur J Pediatr. 2018; 177: 1435-1442
        • Venkatesan T.
        • Levinthal D.J.
        • Tarbell S.E.
        • et al.
        Guidelines on management of cyclic vomiting syndrome in adults by the American Neurogastroenterology and Motility Society and the Cyclic Vomiting Syndrome Association.
        Neurogastroenterol Motil. 2019; 31: e13604
        • Lee L.Y.
        • Abbott L.
        • Mahlangu B.
        • et al.
        The management of cyclic vomiting syndrome: a systematic review.
        Eur J Gastroenterol Hepatol. 2012; 24: 1001-1006
        • Taranukha T.
        • Charan Suresh Kumar V.
        • Seamon A.
        • et al.
        Depression, young age, chronic marijuana use, and interepisodic symptoms predict psychological distress in patients with cyclic vomiting syndrome.
        Neurogastroenterol Motil. 2018; 30: e13245
        • Tarbell S.E.
        • Li B.U.
        Health-related quality of life in children and adolescents with cyclic vomiting syndrome: a comparison with published data on youth with irritable bowel syndrome and organic gastrointestinal disorders.
        J Pediatr. 2013; 163: 493-497
        • Tarbell S.E.
        • Li B.U.
        Anxiety Measures predict health-related quality of life in children and adolescents with cyclic vomiting syndrome.
        J Pediatr. 2015; 167: 633-638.e1
        • Tarbell S.E.
        • Millar A.
        • Laudenslager M.
        • et al.
        Anxiety and physiological responses to the trier social stress test for children in adolescents with cyclic vomiting syndrome.
        Auton Neurosci. 2017; 202: 79-85
        • Ye Z.
        • Xue A.
        • Huang Y.
        • et al.
        Children with cyclic vomiting syndrome: phenotypes, disease burden and mitochondrial DNA analysis.
        BMC Gastroenterol. 2018; 18: 104
        • Levinthal D.J.
        • Romutis S.
        • Rajalaban A.
        • et al.
        Greater intolerance to uncertainty predicts poorer quality of life in adults with cyclic vomiting syndrome.
        Neurogastroenterol Motil. 2021; 33: e14159
        • Fitzpatrick E.
        • Bourke B.
        • Drumm B.
        • et al.
        Outcome for children with cyclical vomiting syndrome.
        Arch Dis Child. 2007; 92: 1001-1004
        • Li B.U.
        • Balint J.P.
        Cyclic vomiting syndrome: evolution in our understanding of a brain-gut disorder.
        Adv Pediatr. 2000; 47: 117-160
        • Foreman M.S.
        • Camp T.
        Cyclic vomiting syndrome.
        Pediatr Rev. 2018; 39: 100-101
        • Wang-Hall J.
        • Li B.U.K.
        • Tarbell S.E.
        Family health-related quality of life in Pediatric cyclic vomiting syndrome.
        J Pediatr Gastroenterol Nutr. 2018; 66: 738-743
        • Deyo R.A.
        • Cherkin D.C.
        • Ciol M.A.
        Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases.
        J Clin Epidemiol. 1992; 45: 613-619
      1. Median weekly earnings of full-time wage and salary workers by detailed occupation and sex. U.S. Bureau of Labor Statistics.
        (Available from:)
        www.bls.gov/cps/aa2018/cpsaat39.htm
        Date: 2019
        Date accessed: March 30, 2020
        • Buono J.L.
        • Carson R.T.
        • Flores N.M.
        Health-related quality of life, work productivity, and indirect costs among patients with irritable bowel syndrome with diarrhea.
        Health Qual Life Outcomes. 2017; 15: 35
        • Lucia-Casadonte C.J.
        • Whaley K.G.
        • Chogle A.S.
        Yield and costs of evaluating children with cyclic vomiting syndrome.
        J Pediatr Gastroenterol Nutr. 2018; 67: 13-17
        • Venkatesan T.
        • Tarbell S.
        • Adams K.
        • et al.
        A survey of emergency department use in patients with cyclic vomiting syndrome.
        BMC Emerg Med. 2010; 10: 4
        • Bhandari S.
        • Venkatesan T.
        Clinical characteristics, comorbidities and hospital outcomes in hospitalizations with cyclic vomiting syndrome: a nationwide analysis.
        Dig Dis Sci. 2017; 62: 2035-2044
        • U.S. Bureau of Labor Statistics
        Full-time / part-time employment. U.S. Department of Labor.
        (Available from:)
        • Parker K.
        Women more than men adjust their careers for family life. Pew Research Center.
        (Available from:)