Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Healthcare resource utilization and direct medical costs associated with index and recurrent Clostridioides difficile infection: a real-world data analysis

Healthcare resource utilization and direct medical costs associated with index and recurrent... JOURNAL OF MEDICAL ECONOMICS 2020, VOL. 23, NO. 6, 603–609 https://doi.org/10.1080/13696998.2020.1724117 Article 0224-RT.R1/1724117 ORIGINAL RESEARCH Healthcare resource utilization and direct medical costs associated with index and recurrent Clostridioides difficile infection: a real-world data analysis a,b c c d,e d c Paul Feuerstadt , Laura Stong , David N. Dahdal , Naomi Sacks , Kathleen Lang and Winnie W. Nelson a b Gastroenterology Center of Connecticut, Hamden, CT, USA; Division of Gastroenterology, Yale University School of Medicine, New Haven, c d CT, USA; Ferring Pharmaceuticals Inc, Parsippany, NJ, USA; Precision Health Economics and Outcomes Research, Boston, MA, USA; Department of Public Health, Tufts University School of Medicine, Boston, MA, USA ABSTRACT ARTICLE HISTORY Received 11 November 2019 Aims: This study aimed to evaluate all-cause economic outcomes, healthcare resource utilization Revised 21 January 2020 (HRU), and costs in patients with Clostridioides difficile infection (CDI) and recurrent CDI (rCDI) using Accepted 24 January 2020 commercial claims from a large database representing various healthcare settings. Materials and methods: A retrospective analysis of commercial claims data from the IQVIA KEYWORDS PharMetrics Plus database was conducted for patients aged 18–64 years with CDI episodes requiring Clostridium difficile infection; inpatient stay with CDI diagnosis code or an outpatient medical claim for CDI plus a CDI treatment. Clostridioides difficile Index CDI episodes occurred between 1 January 2010 and 30 June 2017, including only those where infection (CDI); recurrent patients were observable 6 months before and 12 months after the index episode. Each CDI episode CDI (rCDI); real-world was followed by a 14-d claim-free period. rCDI was defined as another CDI episode within an 8-week outcomes; healthcare resource utilization; direct window following the claim-free period. HRU, all-cause direct medical costs and time to rCDI were cal- medical costs culated over 12 months and stratified by number of rCDI episodes. Results: A total of 46,571 patients with index CDI were included. Mean time from one CDI episode to JEL CLASSIFICATION CODES the next was approximately 1 month. In the 12-month follow-up period, those with no recurrence had I10; I19 1.4 inpatient visits per person and those with 3 or more recurrences had 5.8. Most patients with 3 or more recurrences had 2 or more hospital admissions. The mean annual, total all-cause direct medical costs per patient were $71,980 for those with no recurrence and $207,733 for those with 3 or more recurrences. Limitations: The study included individuals 18–64 years only. A stringent definition of rCDI was used, which may have underestimated the incidence of rCDI. Conclusions: CDI and rCDI are associated with substantial healthcare resource utilization and direct medical costs. Timing of recurrences can be predictable, providing a window of opportunity for inter- ventions. Prevention of multiple rCDI appears essential to reduce healthcare costs. Introduction include previous CDI severity, and presence of hypervirulent 7,8 strain, NAP1/BI/027 . Clostridioides difficile infection (CDI) is the most common The annual economic cost of all CDI in the US is esti- healthcare-associated infection occurring in United States mated to be $5.4 billion, with $4.7 billion of the costs 1,2 (US) hospitals . The Centers for Disease Control and 9 incurred in healthcare settings . Likewise, rCDI is estimated Prevention (CDC) have labelled CDI as a “major health threat” to cost $2.8 billion annually, approximately half of all CDI because of the severity of symptoms, the all-cause mortality costs . Direct medical costs, including inpatient costs, are rate, the potential for antibiotic resistance, and the recur- the main drivers of the overall economic burden of CDI . rence rate in patients who suffer CDI . One means of reducing the overall healthcare burden of CDI It is estimated that approximately 450,000 cases of CDI is to reduce the number of patients who experience occur each year in the US, with increasing incidence over the rCDI episodes. 2,4 last two decades . CDI recurs in approximately 25% of Many of the economic studies may underestimate the patients treated for an initial episode and in up to 40–65% burden of CDI on the healthcare system as the underlying 2,5 of patients who had a prior recurrent CDI (rCDI) . Known data are based on CDI diagnosed and treated only in acute- risk factors for an initial CDI episode include recent systemic care hospitals . The objective of this study was to evaluate antimicrobial therapy, older age (65 years), healthcare economic outcomes, all-cause healthcare resource utilization exposure, and underlying chronic comorbidities, among (HRU), and all-cause costs in patients with CDI and rCDI others . Risk factors for rCDI are largely the same, but also using a large commercial healthcare claims data source CONTACT Winnie Nelson Winnie.nelson@ferring.com Ferring Pharmaceuticals Inc, 100 Interpace Parkway, Parsippany, NJ 07054, USA 2020 Ferring Pharmaceuticals Inc. Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. www.tandfonline.com/ijme 604 P. FEUERSTADT ET AL. All medical claims with a CDI diagnosis that were observed within 14 d of a previous claim with a CDI diagno- sis were considered part of the same episode, as were all days supplied for antibiotic prescriptions, with treatment continuing for the duration of the prescription. Each CDI epi- sode was followed by a 14-d claim-free period after the end Figure 1. Graphical depiction of index CDI and rCDI definitions. of treatment. rCDI was defined as another CDI episode, using the same criteria as above for the index CDI episode, within covering patients treated in different types of healthcare set- an 8-week window following the claim-free period (Figure 1). tings and hospitals. Real-world analysis of clinical outcomes De-identified data were extracted for each individual: age; in patients with CDI and rCDI will be reported in a separ- sex; geographic region; health insurance type; health status ate manuscript. (measured by the Charlson comorbidity index [CCI]); comor- bidities (diabetes, heart disease, cerebrovascular disease, renal disease, liver disease, and prior cancer); baseline CDI Materials and methods risk factors (such as autoimmune conditions, renal insuffi- Study design ciency, and current/history of smoking); and pre-index HRU. This longitudinal, retrospective study used real-world data from the PharMetrics Plus database (IQVIA; Durham, NC), Outcomes which contains claims, enrolment, and demographic data for All outcomes were calculated for the 12-month period after more than 140 million individuals with commercial insurance an index CDI, for all study patients, and by cohorts for num- coverage throughout the US, with coverage of data from ber of rCDI episodes (0 rCDI, 1 rCDI, 2 rCDI, or 3þ rCDI). The over 90% of hospitals and over 90% of all US doctors. number of patients with index CDI and rCDI episodes, and Patients in the PharMetrics Plus database are representative the time to rCDI were calculated. All-cause HRU was calcu- of the national, commercially-insured population in terms of lated, including number and length of inpatient admissions, age and gender for individuals <65 years. The database and number of outpatient visits (including physician office includes all adjudicated medical and prescription drug claims visits, outpatient hospital visits, and emergency department from more than 100 health plans. It includes both inpatient [ED] visits). All-cause direct medical costs were calculated, and outpatient claims, diagnoses and procedures based on including total costs, inpatient costs, outpatient costs (includ- International Classification of Diseases 9th Revision, Clinical ing physician office visits, outpatient hospital visits, ED visits, Modification (ICD-9-CM) and International Classification of and other outpatient services), and prescription drug costs. Diseases, 10th Revision, Clinical Modification (ICD-10-CM), The use of all-cause cost as an outcomes measure is sup- and Current Procedural Terminology (CPT) codes, as well as ported by current methodologic recommendations because retail and mail-order pharmacy claims. it provides a complete picture of cost burden for population 12,13 decision makers . In addition, the available method of Study population using disease codes and drug codes to determine related- ness is not sufficiently reliable to separate disease-related Individuals were included in the study if they were between and disease-unrelated costs. the ages of 18 and 64 years, continuously enrolled, and had at least one inpatient claim with a diagnosis of CDI (ICD-9-CM code: 008.45; ICD-10-CM codes: A04.7, A04.71, or Data analysis A04.72), or one outpatient medical claim with CDI diagnosis Demographic characteristics, costs, and HRU for the relevant code plus a CDI treatment. Treatment was defined as a cohorts were displayed using counts and percentages for prescription for an antibiotic used to treat CDI (vancomycin, categorical variables and measures of central tendency fidaxomicin, metronidazole, or rifaximin), bezlotoxumab (mean [standard deviation, SD]) for continuous variables. (monoclonal antibody therapy), or faecal microbiota trans- Multivariate logistic regression, adjusting for age, sex, illness plant (FMT). FMT was identified by CPT codes 44705 or burden (CCI), geographic region, health plan type, pre-index 44799, or Healthcare Common Procedure Coding System medication use (pre-index gastric acid-suppressing agents, (HCPCS) code G0455. antibiotics), pre-index medical procedures and treatments Index CDI episodes occurred between 1 January 2010 and 30 June 2017 (Figure 1). Only patients who were observable (transplant, gastrointestinal surgery, enteral feeding, chemo- 6 months before and 12 months after the index CDI episode therapy), pre-index inpatient admission, and immunosup- pressed status (Y/N), was conducted to evaluate the odds of were included. The pre-index period was used to quantify pre-CDI healthcare exposure and to indicate that the first multiple recurrences. observed CDI diagnosis was not a recurrent episode, while Costs were converted to 2018 dollars using the medical the post-index requirement allowed sufficient time for care component of the Consumer Price Index . All statistical analyses were conducted with SAS version 9.3 (SAS Institute, observing one or more recurrences following the index diagnosis. Inc., Cary, NC). JOURNAL OF MEDICAL ECONOMICS 605 patient also was highest for those with three or more Results rCDI episodes. A total of 46,571 patients with an index CDI episode were The mean annual, total all-cause direct medical costs per included, with 3,129 (6.7%) having 1 recurrence, 472 (1.0%) patient within the 12-month follow-up period were substan- having 2 recurrences, and 134 (0.3%) having 3 or more tially different by rCDI group, from almost $72,000 for those recurrences (Table 1). A larger proportion of patients were with no recurrence to $207,700 for those with 3 or more female and most had a preferred provider organization recurrences (Figure 4). Inpatient costs were the key cost (PPO) health plan. The mean age was consistent across driver, accounting for 61–70% of the total costs across the recurrence groups, at approximately 48 years. study cohorts. Outpatient costs comprised outpatient hos- Geographically, the lowest rates of CDI and rCDI occurred pital visits, physician office visits, ED visits, and other out- in the West, and the highest rates of rCDI were seen in the patient services (including laboratory and imaging tests). Of Midwest. The mean (SD) baseline CCI score was lowest for the total outpatient costs, outpatient hospital visits those with no recurrences [1.2 (1.9)] and highest for those accounted for the majority of the cost (56–58% of total): with 3 or more recurrences [2.3 (2.5)]. $12,111; $18,516; $22,542; and $25,828 by increasing recur- The mean time from one CDI episode to a recurrence rence group. ED costs were, by increasing recurrence group, (within the 8-week window, per definition) was consistent at $1,329; $2,127; $3,237; and $4,633. approximately 1 month (Figure 2). There were no substantial Through multivariate logistic regression, there was only 1 differences in the time to next CDI episode for patients who covariate that was significantly associated with having 2 or had 1 recurrence, 2 recurrences, or 3 or more recurrences more rCDI episodes vs. 1 rCDI (Table 3). Those with pre-index (values ranged from 30.7 to 33.3 d). medical procedures and treatments, including transplant, During the 12-month follow-up period, HRU was reflective gastrointestinal surgery, enteral feeding, or chemotherapy, of the number of rCDI episodes. The number of inpatient vis- were significantly more likely (odds ratio 1.27; 95% confi- its per person and ED visits per person was highest for those dence interval: 1.04, 1.55) to have 2 or more recurrences with the greatest number of rCDI episodes (Figure 3). The than patients without pre-index medical procedures (p¼.02). majority of patients (69.4%) who had 3 or more recurrences had at least 2 hospital admissions during the follow-up All other patient and clinical characteristics were not signifi- cantly associated with having more than 1 episode of rCDI, period, a figure which was lower for patients with indicating the similarities between these cohorts, although a fewer recurrences (62.1% for 2 rCDI, 53.0% for 1 rCDI, and 29.1% for 0 rCDI) (Table 2). The number of prescriptions per number of characteristics were borderline significant. Table 1. Demographic and baseline characteristics. No recurrence 1 Recurrence 2 Recurrences 3þ Recurrences N¼ 42,836 N¼ 3,129 N¼ 472 N¼ 134 Age, years, mean (SD) 47.4 (12.7) 48.3 (12.8) 47.9 (13.0) 48.7 (11.5) Female, n (%) 26,625 (62.2) 2,036 (65.1) 319 (67.6) 82 (61.2) Geographic region, n (%) Midwest 13,190 (30.8) 981 (31.4) 147 (31.1) 33.6 (45) Northeast 9,741 (22.7) 786 (25.1) 133 (28.2) 42 (31.3) South 14,585 (34.1) 958 (30.6) 140 (29.7) 33 (24.6) West 4,663 (10.9) 360 (11.5) 51 (10.8) 12 (9.0) Unknown 657 (1.5) 44 (1.4) Health plan, n (%) PPO 32,990 (77.0) 2,347 (75.0) 344 (72.9) 84 (62.7) HMO 6,103 (14.3) 519 (16.6) 87 (18.4) 36 (26.9) CDHP 269 (0.6) 16 (0.5) Other 3,266 (7.6) 233 (7.5) 34 (7.2) 12 (9.0) Unknown 208 (0.5) 14 (0.5) CCI score, mean (SD) 1.2 (1.9) 1.5 (2.2) 1.8 (2.3) 2.3 (2.5) Comorbid condition, n (%) Autoimmune disease 7,745 (18.1) 723 (23.1) 116 (24.6) 53 (39.6) Ulcerative colitis 2,326 (5.4) 238 (7.6) 39 (8.3) 21 (15.7) Crohn’s disease 1,782 (4.2) 175 (5.6) 22 (4.7) 11 (8.2) Type 1 diabetes 1,359 (3.2) 134 (4.3) 18 (3.8) 11 (8.2) Renal insufficiency 5,618 (13.1) 571 (18.3) 105 (22.3) 36 (26.9) Pre-index medications, n (%) Gastric acid-suppressing agents 11,943 (27.9) 1,028 (32.9) 184 (39.0) 51 (38.1) Antibiotics 33,411 (78.0) 2,509 (80.2) 381 (80.7) 103 (76.9) Pre-index healthcare use, n (%) Inpatient admission 13,938 (32.5) 1,307 (41.8) 236 (50.0) 81 (60.5) Outpatient hospital visit 32,584 (76.1) 2,576 (82.3) 404 (85.6) 116 (86.6) ED visit 19,534 (45.6) 1,581 (50.5) 268 (56.8) 77 (57.5) Outpatient office visit 40,064 (93.5) 2,951 (94.3) 455 (96.4) 129 (96.3) For patient privacy reasons and consistent with data reporting practices for the Centers for Medicare and Medicaid Services, data are not shown for cells in which the sample size was 10. Abbreviations. CCI, Charlson comorbidity index; CDHP, consumer-directed health plan; ED, emergency department; HMO, health maintenance organization; PPO, preferred provider organization; SD, standard deviation. 606 P. FEUERSTADT ET AL. 540 patients, the total hospitalization cost of treating a rCDI Discussion episode was 2.2 times that of primary CDI . Our study pro- CDI and rCDI are associated with substantial all-cause HRU vides a broader picture for CDI and rCDI medical costs, in and direct medical costs. Mean total, all-cause, direct medical more current dollars, that are not exclusively related to the costs for patients with rCDI varied by number of rCDI epi- CDI episode/hospitalization. Our study also focuses on a sodes and ranged between $131,000 for patients with 1 younger cohort than most other studies, with included recurrence to more than $200,000 in patients with 3 or more patients between the ages of 18 and 64 years. recurrences in the 12 months following an index episode; After a 6-month baseline period that was devoid of any costs for inpatient stays made up the majority of direct costs CDI claims, the average time from the index CDI episode to for all cohorts. The mean time between rCDI episodes was each rCDI episode was approximately 1 month, similar to stable at about 1 month from the previous episode, with a what has been reported previously . This likely represents standard deviation of approximately 15 d regardless of study the highest risk period for recurrence within the “window of cohort, indicating that the recurrence timing in this cohort vulnerability.” Once the colonic microbiota has been signifi- aged 18–64 years seems predictable. cantly depleted, the antibiotic treatment suppresses the CDI, Prior studies have shown a mean CDI-attributable cost but subsequently there might be insufficient time to restore from $8,911 to $30,049 per patient/admission/episode/ the microbiota to a state where it is protective against reacti- infection (2014 US dollars) . In one single centre study of vation of the latent C. difficile spores. Therefore, rate of recur- rence and time to recurrence could be expected to be similar across episodes, without the presence of other influ- encing factors (e.g. concomitant antibiotics). Because the time to recurrence was relatively brief, we would expect the patient’s health plan at the time of the index CDI to assume most of the costs associated with the subsequent rCDI epi- sode(s). This 1-month period also presents itself as a window of opportunity for interventions that can reduce the risk for future rCDI episodes. The findings of the logistic regression are important particularly because minimally significant fac- Figure 2. The mean (SD) days from one CDI episode to a recurrence (within tors were found that predict multiple recurrences vs. only 1 the 8-week window defining rCDI) was consistent at approximately 1 month and did not depend on the number of rCDI episodes (n¼ 3,735 patients). recurrence. Therefore, it seems that preventing any recur- rence is the key objective, which should be an imperative for all CDI care. Though there are published analyses reporting risk factors for any CDI recurrence, the risk factors for mul- tiple rCDI vs. 1 rCDI have not been extensively studied 8,18–21 before . HRU was high for all patients with an index CDI, with the highest utilization for those with 3 or more recurrences. Other analyses of HRU have shown a similar hospital length of stay of 8.0 d for the index CDI episode, and 9.3 d for a rCDI episode . In 2016, the mean hospital length of stay in the US for all conditions was 4.6 d, indicating that CDI places a higher burden on HRU than the average admission .A single centre study showed similar rates of ICU use to our study, with 9.4% of patients with rCDI having an inpatient admission with ICU stay . An analysis of electronic records from 85 hospitals in the US showed the highest incidence of Figure 3. In the 12-month follow-up period, the mean number of visits per 25 community-onset CDI was in the Northeast and Midwest . patient for the emergency department (ED) or for an inpatient admission was Our results support some of the earlier studies, with the highest for those with 3 or more CDI recurrences, at 4.6 visits and 5.8 visits, respectively. highest rates of rCDI occurring in the Midwest. Table 2. All-cause healthcare resource utilization during 12-month follow-up. No recurrence 1 Recurrence 2 Recurrences 3þ Recurrences Resource utilized (N¼ 42,836) (N¼ 3,129) (N¼ 472) (N¼ 134) Patients with admissions, n (%) 0 Admissions 15,875 (37.1) 898 (28.7) 111 (23.5) 29 (21.6) 1 Admission 14,490 (33.8) 572 (18.3) 68 (14.4) 12 (9.0) 2 Admissions 12,471 (29.1) 1,659 (53.0) 293 (62.1) 93 (69.4) LOS per admission, mean (SD) 7.7 (9.7) 8.3 (7.8) 7.9 (6.2) 8.5 (6.6) Inpatient admission with an ICU Stay, n (%) 2,332 (5.4) 251 (8.0) 27 (5.7) 17 (12.7) Prescriptions per patient, mean (SD) 40.5 (39.5) 51.5 (42.6) 61.6 (46.4) 65.0 (51.0) Abbreviations. ICU, intensive care unit; LOS, length of stay; SD, standard deviation. JOURNAL OF MEDICAL ECONOMICS 607 burden of rCDI, and, therefore, we believe, a strict definition increased the validity of the findings. There may, unfortu- nately, have been cases of recurrent CDI that did not meet our case definition; thus, the reported incidence of rCDI may be an underestimation. The PharMetrics Plus database contains adjudicated claims from payers for a diverse US population. The cost data col- lected reflect the payments made by the health plans or employers, therefore, we believe this data source is suffi- ciently accurate and valid to quantify the economic impact of rCDI to US commercial payers. As the data are solely from the US, they may not be generalizable to other countries. In this study, it is likely that the medical costs for the cohort with 3 or more recurrences are underestimated, because the cost summation stopped at 12 months following the index Figure 4. Total, all-cause, direct medical costs during the 12-month period case of CDI. It is most likely that those patients who suffered after an index CDI episode were substantial for any patient with recurrent CDI, and highest for those with 3 or more recurrences. Inpatient costs made up the 3 or more recurrences incurred additional CDI-related costs majority of the total costs. Costs are adjusted to 2018 dollars. beyond 12 months from the index episode. One limitation of this study is that it did not include indi- Table 3. Multivariate logistic regression for 2þ recurrences vs. 1 recurrence. viduals 65 years, as almost all of these individuals in the Odds ratio 95% CI p Value US are covered by Medicare, and the database used for the Female (vs. male) 1.13 0.93 1.36 .21 analysis contained only commercial claims. Previous research Age, years (vs.18–39) showed that older patients are more susceptible to rCDI, and 40–51 0.99 0.77 1.27 .92 52–58 1.04 0.81 1.34 .76 the per-patient cost of treatment could be higher due to the 59þ 0.78 0.60 1.02 .07 9,26 impact of chronic conditions . Future studies focusing on Geographic region (vs. Midwest) Northeast 1.07 0.84 1.35 .59 the older, frail population would add to the current body of South 0.91 0.72 1.14 .41 evidence. Also, a retrospective analysis of insurance claims West 0.88 0.64 1.21 .42 data is subject to potential miscoding, however, these data Type of benefit plan (vs. PPO) HMO 1.26 0.99 1.59 .06 errors are expected to be random and equal across the HDHP/CDHP 2.17 0.87 5.45 .10 population. There may be missing encounter data when Other 1.07 0.76 1.51 .68 patients paid out of pocket, which would lead to an under- Unknown 1.03 0.29 3.69 .96 CCI Score (vs.0) estimation of costs. To address the key study objective, the 1 1.22 0.92 1.61 .17 study included patients who had a minimum of 18 months 2þ 1.24 0.98 1.55 .07 of continuous enrolment. These inclusion criteria would have Baseline medications Gastric acid suppressing agents (vs. not) 1.14 0.94 1.38 .17 excluded patients who disenroled sooner, including patients Antibiotics (vs. not) 1.00 0.80 1.25 .99 who died due to CDI, rCDI, or other reasons. In addition, Immunosuppressed (vs. not) 1.16 0.95 1.41 .16 individuals were excluded from this study if they lost com- Medical procedures and treatments (vs. none) 1.27 1.04 1.55 .02 Year of index CDI (vs. 2010–2013) mercial health insurance for any reason during the study 2014–2017 1.10 0.92 1.31 .32 period and could not be followed for 12 months; therefore, Inpatient admission in pre-index period (vs. none) 1.23 0.99901.52 .051 patients who became too disabled to work may have been Immunosuppressed defined as taking immunosuppressant agent(s) and/or excluded by this criterion. The impact of this limitation might has autoimmune disease during baseline period. Abbreviations. CCI, Charlson comorbidity index; CDHP, consumer-directed have led to an under-estimation of the resource use and health plan; CDI, Clostridioides difficile infection; CI, confidence interval; HMO, cost findings. The database does not define whether the health maintenance organization; PPO, preferred provider organization. patients had healthcare vs. community-acquired infection. The recurrence rates seen in this study are lower that Because this commercial database does not contain mortality what has been reported elsewhere (between 13.5% and 35% data for patient privacy protection, the analysis cannot dis- 2,26 recurring within 30 d after initial CDI diagnosis) . These cern the reason for disenrolment. As such, the study was lower rates are likely due to our study including a younger designed to focus on depicting economic burden, while (18–64 years) cohort than those who are most frequently unable to report mortality burden of rCDI. 9,26,27 affected by CDI (65 years) . Younger patients tend to have fewer medical comorbidities, less requirement for con- Conclusion comitant antimicrobials, and likely lower rates of CDI recur- rence. The mean CCI values reported here for patients with CDI and rCDI are associated with substantial all-cause HRU any rCDI are slightly lower than what was reported in other and direct medical costs. As expected, repeat recurrences 28–31 observational and randomized, controlled trials in CDI . multiply the economic burden further. Despite a lack of pre- Another potential reason for our lower observed recurrence dictors for multiple vs. 1 rCDI from patient characteristics, the rates was our stringent criteria identifying likely rCDI cases. timing of rCDI recurrence appears predictable. Leveraging The study’s key objective was to quantify the economic this window of opportunity for early interventions that aim 608 P. FEUERSTADT ET AL. [8] Zilberberg MD, Reske K, Olsen M, et al. Risk factors for recurrent at reducing recurrences is key to addressing the economic Clostridium difficile infection (CDI) hospitalization among hospital- burden of rCDI. Future research could consider a prospective ized patients with an initial CDI episode: a retrospective cohort study design of rCDI costs and consider a cohort study. BMC Infect Dis. 2014;14(1):306. aged 65 years. [9] Desai K, Gupta SB, Dubberke ER, et al. Epidemiological and economic burden of Clostridium difficile in the United States: estimates from a modeling approach. BMC Infect Dis. 2016;16(1): Transparency [10] Rodrigues R, Barber GE, Ananthakrishnan AN. A comprehensive Declaration of funding study of costs associated with recurrent Clostridium difficile infec- tion. Infect Control Hosp Epidemiol. 2017;38(2):196–202. The study was funded by Ferring Pharmaceuticals Inc. (Parsippany, NJ). [11] Dubberke ER, Olsen MA. Burden of Clostridium difficile on the healthcare system. Clin Infect Dis. 2012;55(2):S88–S92. [12] van Baal P, Morton A, Meltzer D, et al. Future unrelated medical Declaration of financial/other interests costs need to be considered in cost effectiveness analysis. Eur J Health Econ. 2019;20(1):1–5. Drs. Stong, Nelson, and Dahdal are employees of Ferring [13] Rappange DR, van Baal PHM, van Exel NJA, et al. Unrelated Pharmaceuticals Inc. Dr. Feuerstadt has served as a consultant to and on medical costs in life-years gained: should they be included in eco- the speaker’s bureau for Merck and Co, and has served as a consultant nomic evaluations of healthcare interventions? Pharmacoeconomics. for Ferring Pharmaceuticals Inc. and Roche Pharmaceuticals. Drs. Sacks 2008;10:815–830. and Lang are employees of Precision Health Economic and Outcomes [14] U.S. Bureau of Labor Statistics. Measuring price change in the Research and provided consulting services to Ferring CPI: medical care [Internet]; [cited 2020 Jan 9]. Available from: Pharmaceuticals Inc. https://www.bls.gov/cpi/factsheets/medical-care.htm A peer reviewer on this manuscript has disclosed that they have [15] Nanwa N, Kendzerska T, Krahn M, et al. The economic impact of received consultancy fees from Astellas Pharma. The peer reviewers on Clostridium difficile infection: a systematic review. Am J this manuscript have no other relevant financial relationships or other- Gastroenterol. 2015;110(4):511–519. wise to disclose. [16] Shah DN, Aitken SL, Barragan LF, et al. Economic burden of pri- mary compared with recurrent Clostridium difficile infection in hospitalized patients: a prospective cohort study. J Hosp Infect. Acknowledgements 2016;93(3):286–289. [17] Kelly CP. Can we identify patients at high risk of recurrent Medical writing and editorial support was provided by Agnella Izzo Clostridium difficile infection? Clin Microbiol Infect. 2012;18:21–27. Matic, PhD, CMPP (AIM Biomedical, LLC) and was funded by Ferring [18] Abou Chakra CN, Pepin J, Sirard S, et al. Risk factors for recur- Pharmaceuticals Inc. rence, complications an mortality in Clostridium difficile infection: a systematic review. PLoS One. 2014;9(6):e98400. [19] Escobar GJ, Baker JM, Kipnis P, et al. Prediction of recurrenct Author contributions Clostridium difficile infection using comprehensive electronic med- LS, WN, DND, NS, and KL designed and conducted the study. All authors ical records in an integrated healthcare delivery system. Infect analyzed and interpreted the data, drafted and critically revised the art- Control Hosp Epidemiol. 2017;38(10):1196–1203. icle for important intellectual content, and approved the article for [20] LaBarbera FD, Nikiforov I, Parvathenani A, et al. A prediction publication. model for Clostridium difficile recurrence. J Comm Hosp Intern Med Perspect. 2015;5(1):26033. [21] van Beurden YH, Nezami S, Mulder CJJ, et al. Host factors are more important in predicting recurrent Clostridium difficile infec- Previous presentations tion than ribotype and use of antibiotics. Clin Microbiol Infect. Portions of the data contained in this manuscript appeared in abstract/ 2018;24(1):85.e1–85.e4. poster form at IDWeek 2019, October 2-6, 2019, Poster 2374. This manu- [22] Zhang D, Prabhu VS, Marcella SW. Attributable healthcare script has not been submitted and is not under consideration for publi- resource utilization and costs for patients with primary and recur- cation anywhere else. rent Clostridium difficile infection in the United States. Clin Infect Dis. 2018;66(9):1326–1332. [23] Freeman WJ, Weiss AJ, Heslin KC. Overview of U.S. hospital stays in References 2016: variation by geographic region. HCUP Statistical Brief #246. December 2018. Agency for Healthcare Research and Quality, [1] Ofosu A. Clostridium difficile infection: a review of current and Rockville, MD. www.hcup-us.ahrq.gov/reports/statbriefs/sb246- emerging therapies. Ann Gastroenterol. 2016;29(2):147–154. Geographic-Variation-Hospital-Stays.pdf. [2] Smits WK, Lyras D, Lacy DB, et al. Clostridium difficile infection. [24] Aitken SL, Joseph TB, Shah DN, et al. Healthcare resource utiliza- Nat Rev Dis Primers. 2016;2:16020. tion for recurrence Clostridium difficile infection in a large univer- [3] Centers for Disease Control and Prevention. Clostridioides difficile sity hospital in Houston, Texas. PLoS One. 2014;9(7):e102848. infection [Internet]; [cited 2019 Aug 13]. Available from: https:// [25] Zilberberg MD, Tabak YP, Sievert DM, et al. Using electronic www.cdc.gov/hai/organisms/cdiff/cdiff_infect.html health information to risk-stratify rates of Clostridium difficile [4] Reveles KR, Lee GC, Boyd NK, et al. The rise in Clostridium difficile infection in US hospitals. Infect Control Hosp Epidemiol. 2011; infection incidence among hospitalized adults in the United 32(7):649–655. States: 2001–2010. Am J Infect Control. 2014;42(10):1028–1032. [26] Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile [5] Leffler DA, Lamont JT. Clostridium difficile infection. N Engl J Med. infection in the United States. N Engl J Med. 2015;372(9): 2015;372(16):1539–1548. 825–834. [6] DePestel DD, Aronoff DM. Epidemiology of Clostridium difficile [27] Pechal A, Lin K, Allen S, et al. National age group trends in infection. J Pharm Pr. 2013;26(5):464–475. Clostridium difficile infection incidence and health outcomes [7] Song JH, Kim YS. Recurrent Clostridium difficile infection: risk fac- in United States community hospitals. BMC Infect Dis. 2016;16(1): tors, treatment, and prevention. Gut Liver. 2019;13(1):16–24. 682. JOURNAL OF MEDICAL ECONOMICS 609 [28] Cammarota G, Masucci L, Ianiro G, et al. Randomised clinical trial: [30] Delholm-Lambertsen E, Hall BK, Jørgensen SMD, et al. Cost sav- faecal microbiota transplantation by colonoscopy vs. vancomycin ings following faecal microbiota transplantation for recurrent for the treatment of recurrent Clostridium difficile infection. Clostridium difficile infection. Ther Adv Gastroenterol. 2019;12: Aliment Pharmacol Ther. 2015;41(9):835–843. 1–14. [29] Charlson ME, Pompei P, Ales KL, et al. A new method of classify- [31] McFarland LV, Surawicz CM, Rubin M, et al. Recurrent Clostridium ing prognostic comorbidity in longitudinal studies: development difficile disease: epidemiology and clinical characteristics. Infect and validation. J Chron Dis. 1987;40(5):373–383. Control Hosp Epidemiol. 1999;20(01):43–50. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Medical Economics Taylor & Francis

Healthcare resource utilization and direct medical costs associated with index and recurrent Clostridioides difficile infection: a real-world data analysis

Loading next page...
 
/lp/taylor-francis/healthcare-resource-utilization-and-direct-medical-costs-associated-2WsDIHyTUv

References (31)

Publisher
Taylor & Francis
Copyright
© 2020 Ferring Pharmaceuticals Inc. Published by Informa UK Limited, trading as Taylor & Francis Group.
ISSN
1941-837X
eISSN
1369-6998
DOI
10.1080/13696998.2020.1724117
Publisher site
See Article on Publisher Site

Abstract

JOURNAL OF MEDICAL ECONOMICS 2020, VOL. 23, NO. 6, 603–609 https://doi.org/10.1080/13696998.2020.1724117 Article 0224-RT.R1/1724117 ORIGINAL RESEARCH Healthcare resource utilization and direct medical costs associated with index and recurrent Clostridioides difficile infection: a real-world data analysis a,b c c d,e d c Paul Feuerstadt , Laura Stong , David N. Dahdal , Naomi Sacks , Kathleen Lang and Winnie W. Nelson a b Gastroenterology Center of Connecticut, Hamden, CT, USA; Division of Gastroenterology, Yale University School of Medicine, New Haven, c d CT, USA; Ferring Pharmaceuticals Inc, Parsippany, NJ, USA; Precision Health Economics and Outcomes Research, Boston, MA, USA; Department of Public Health, Tufts University School of Medicine, Boston, MA, USA ABSTRACT ARTICLE HISTORY Received 11 November 2019 Aims: This study aimed to evaluate all-cause economic outcomes, healthcare resource utilization Revised 21 January 2020 (HRU), and costs in patients with Clostridioides difficile infection (CDI) and recurrent CDI (rCDI) using Accepted 24 January 2020 commercial claims from a large database representing various healthcare settings. Materials and methods: A retrospective analysis of commercial claims data from the IQVIA KEYWORDS PharMetrics Plus database was conducted for patients aged 18–64 years with CDI episodes requiring Clostridium difficile infection; inpatient stay with CDI diagnosis code or an outpatient medical claim for CDI plus a CDI treatment. Clostridioides difficile Index CDI episodes occurred between 1 January 2010 and 30 June 2017, including only those where infection (CDI); recurrent patients were observable 6 months before and 12 months after the index episode. Each CDI episode CDI (rCDI); real-world was followed by a 14-d claim-free period. rCDI was defined as another CDI episode within an 8-week outcomes; healthcare resource utilization; direct window following the claim-free period. HRU, all-cause direct medical costs and time to rCDI were cal- medical costs culated over 12 months and stratified by number of rCDI episodes. Results: A total of 46,571 patients with index CDI were included. Mean time from one CDI episode to JEL CLASSIFICATION CODES the next was approximately 1 month. In the 12-month follow-up period, those with no recurrence had I10; I19 1.4 inpatient visits per person and those with 3 or more recurrences had 5.8. Most patients with 3 or more recurrences had 2 or more hospital admissions. The mean annual, total all-cause direct medical costs per patient were $71,980 for those with no recurrence and $207,733 for those with 3 or more recurrences. Limitations: The study included individuals 18–64 years only. A stringent definition of rCDI was used, which may have underestimated the incidence of rCDI. Conclusions: CDI and rCDI are associated with substantial healthcare resource utilization and direct medical costs. Timing of recurrences can be predictable, providing a window of opportunity for inter- ventions. Prevention of multiple rCDI appears essential to reduce healthcare costs. Introduction include previous CDI severity, and presence of hypervirulent 7,8 strain, NAP1/BI/027 . Clostridioides difficile infection (CDI) is the most common The annual economic cost of all CDI in the US is esti- healthcare-associated infection occurring in United States mated to be $5.4 billion, with $4.7 billion of the costs 1,2 (US) hospitals . The Centers for Disease Control and 9 incurred in healthcare settings . Likewise, rCDI is estimated Prevention (CDC) have labelled CDI as a “major health threat” to cost $2.8 billion annually, approximately half of all CDI because of the severity of symptoms, the all-cause mortality costs . Direct medical costs, including inpatient costs, are rate, the potential for antibiotic resistance, and the recur- the main drivers of the overall economic burden of CDI . rence rate in patients who suffer CDI . One means of reducing the overall healthcare burden of CDI It is estimated that approximately 450,000 cases of CDI is to reduce the number of patients who experience occur each year in the US, with increasing incidence over the rCDI episodes. 2,4 last two decades . CDI recurs in approximately 25% of Many of the economic studies may underestimate the patients treated for an initial episode and in up to 40–65% burden of CDI on the healthcare system as the underlying 2,5 of patients who had a prior recurrent CDI (rCDI) . Known data are based on CDI diagnosed and treated only in acute- risk factors for an initial CDI episode include recent systemic care hospitals . The objective of this study was to evaluate antimicrobial therapy, older age (65 years), healthcare economic outcomes, all-cause healthcare resource utilization exposure, and underlying chronic comorbidities, among (HRU), and all-cause costs in patients with CDI and rCDI others . Risk factors for rCDI are largely the same, but also using a large commercial healthcare claims data source CONTACT Winnie Nelson Winnie.nelson@ferring.com Ferring Pharmaceuticals Inc, 100 Interpace Parkway, Parsippany, NJ 07054, USA 2020 Ferring Pharmaceuticals Inc. Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. www.tandfonline.com/ijme 604 P. FEUERSTADT ET AL. All medical claims with a CDI diagnosis that were observed within 14 d of a previous claim with a CDI diagno- sis were considered part of the same episode, as were all days supplied for antibiotic prescriptions, with treatment continuing for the duration of the prescription. Each CDI epi- sode was followed by a 14-d claim-free period after the end Figure 1. Graphical depiction of index CDI and rCDI definitions. of treatment. rCDI was defined as another CDI episode, using the same criteria as above for the index CDI episode, within covering patients treated in different types of healthcare set- an 8-week window following the claim-free period (Figure 1). tings and hospitals. Real-world analysis of clinical outcomes De-identified data were extracted for each individual: age; in patients with CDI and rCDI will be reported in a separ- sex; geographic region; health insurance type; health status ate manuscript. (measured by the Charlson comorbidity index [CCI]); comor- bidities (diabetes, heart disease, cerebrovascular disease, renal disease, liver disease, and prior cancer); baseline CDI Materials and methods risk factors (such as autoimmune conditions, renal insuffi- Study design ciency, and current/history of smoking); and pre-index HRU. This longitudinal, retrospective study used real-world data from the PharMetrics Plus database (IQVIA; Durham, NC), Outcomes which contains claims, enrolment, and demographic data for All outcomes were calculated for the 12-month period after more than 140 million individuals with commercial insurance an index CDI, for all study patients, and by cohorts for num- coverage throughout the US, with coverage of data from ber of rCDI episodes (0 rCDI, 1 rCDI, 2 rCDI, or 3þ rCDI). The over 90% of hospitals and over 90% of all US doctors. number of patients with index CDI and rCDI episodes, and Patients in the PharMetrics Plus database are representative the time to rCDI were calculated. All-cause HRU was calcu- of the national, commercially-insured population in terms of lated, including number and length of inpatient admissions, age and gender for individuals <65 years. The database and number of outpatient visits (including physician office includes all adjudicated medical and prescription drug claims visits, outpatient hospital visits, and emergency department from more than 100 health plans. It includes both inpatient [ED] visits). All-cause direct medical costs were calculated, and outpatient claims, diagnoses and procedures based on including total costs, inpatient costs, outpatient costs (includ- International Classification of Diseases 9th Revision, Clinical ing physician office visits, outpatient hospital visits, ED visits, Modification (ICD-9-CM) and International Classification of and other outpatient services), and prescription drug costs. Diseases, 10th Revision, Clinical Modification (ICD-10-CM), The use of all-cause cost as an outcomes measure is sup- and Current Procedural Terminology (CPT) codes, as well as ported by current methodologic recommendations because retail and mail-order pharmacy claims. it provides a complete picture of cost burden for population 12,13 decision makers . In addition, the available method of Study population using disease codes and drug codes to determine related- ness is not sufficiently reliable to separate disease-related Individuals were included in the study if they were between and disease-unrelated costs. the ages of 18 and 64 years, continuously enrolled, and had at least one inpatient claim with a diagnosis of CDI (ICD-9-CM code: 008.45; ICD-10-CM codes: A04.7, A04.71, or Data analysis A04.72), or one outpatient medical claim with CDI diagnosis Demographic characteristics, costs, and HRU for the relevant code plus a CDI treatment. Treatment was defined as a cohorts were displayed using counts and percentages for prescription for an antibiotic used to treat CDI (vancomycin, categorical variables and measures of central tendency fidaxomicin, metronidazole, or rifaximin), bezlotoxumab (mean [standard deviation, SD]) for continuous variables. (monoclonal antibody therapy), or faecal microbiota trans- Multivariate logistic regression, adjusting for age, sex, illness plant (FMT). FMT was identified by CPT codes 44705 or burden (CCI), geographic region, health plan type, pre-index 44799, or Healthcare Common Procedure Coding System medication use (pre-index gastric acid-suppressing agents, (HCPCS) code G0455. antibiotics), pre-index medical procedures and treatments Index CDI episodes occurred between 1 January 2010 and 30 June 2017 (Figure 1). Only patients who were observable (transplant, gastrointestinal surgery, enteral feeding, chemo- 6 months before and 12 months after the index CDI episode therapy), pre-index inpatient admission, and immunosup- pressed status (Y/N), was conducted to evaluate the odds of were included. The pre-index period was used to quantify pre-CDI healthcare exposure and to indicate that the first multiple recurrences. observed CDI diagnosis was not a recurrent episode, while Costs were converted to 2018 dollars using the medical the post-index requirement allowed sufficient time for care component of the Consumer Price Index . All statistical analyses were conducted with SAS version 9.3 (SAS Institute, observing one or more recurrences following the index diagnosis. Inc., Cary, NC). JOURNAL OF MEDICAL ECONOMICS 605 patient also was highest for those with three or more Results rCDI episodes. A total of 46,571 patients with an index CDI episode were The mean annual, total all-cause direct medical costs per included, with 3,129 (6.7%) having 1 recurrence, 472 (1.0%) patient within the 12-month follow-up period were substan- having 2 recurrences, and 134 (0.3%) having 3 or more tially different by rCDI group, from almost $72,000 for those recurrences (Table 1). A larger proportion of patients were with no recurrence to $207,700 for those with 3 or more female and most had a preferred provider organization recurrences (Figure 4). Inpatient costs were the key cost (PPO) health plan. The mean age was consistent across driver, accounting for 61–70% of the total costs across the recurrence groups, at approximately 48 years. study cohorts. Outpatient costs comprised outpatient hos- Geographically, the lowest rates of CDI and rCDI occurred pital visits, physician office visits, ED visits, and other out- in the West, and the highest rates of rCDI were seen in the patient services (including laboratory and imaging tests). Of Midwest. The mean (SD) baseline CCI score was lowest for the total outpatient costs, outpatient hospital visits those with no recurrences [1.2 (1.9)] and highest for those accounted for the majority of the cost (56–58% of total): with 3 or more recurrences [2.3 (2.5)]. $12,111; $18,516; $22,542; and $25,828 by increasing recur- The mean time from one CDI episode to a recurrence rence group. ED costs were, by increasing recurrence group, (within the 8-week window, per definition) was consistent at $1,329; $2,127; $3,237; and $4,633. approximately 1 month (Figure 2). There were no substantial Through multivariate logistic regression, there was only 1 differences in the time to next CDI episode for patients who covariate that was significantly associated with having 2 or had 1 recurrence, 2 recurrences, or 3 or more recurrences more rCDI episodes vs. 1 rCDI (Table 3). Those with pre-index (values ranged from 30.7 to 33.3 d). medical procedures and treatments, including transplant, During the 12-month follow-up period, HRU was reflective gastrointestinal surgery, enteral feeding, or chemotherapy, of the number of rCDI episodes. The number of inpatient vis- were significantly more likely (odds ratio 1.27; 95% confi- its per person and ED visits per person was highest for those dence interval: 1.04, 1.55) to have 2 or more recurrences with the greatest number of rCDI episodes (Figure 3). The than patients without pre-index medical procedures (p¼.02). majority of patients (69.4%) who had 3 or more recurrences had at least 2 hospital admissions during the follow-up All other patient and clinical characteristics were not signifi- cantly associated with having more than 1 episode of rCDI, period, a figure which was lower for patients with indicating the similarities between these cohorts, although a fewer recurrences (62.1% for 2 rCDI, 53.0% for 1 rCDI, and 29.1% for 0 rCDI) (Table 2). The number of prescriptions per number of characteristics were borderline significant. Table 1. Demographic and baseline characteristics. No recurrence 1 Recurrence 2 Recurrences 3þ Recurrences N¼ 42,836 N¼ 3,129 N¼ 472 N¼ 134 Age, years, mean (SD) 47.4 (12.7) 48.3 (12.8) 47.9 (13.0) 48.7 (11.5) Female, n (%) 26,625 (62.2) 2,036 (65.1) 319 (67.6) 82 (61.2) Geographic region, n (%) Midwest 13,190 (30.8) 981 (31.4) 147 (31.1) 33.6 (45) Northeast 9,741 (22.7) 786 (25.1) 133 (28.2) 42 (31.3) South 14,585 (34.1) 958 (30.6) 140 (29.7) 33 (24.6) West 4,663 (10.9) 360 (11.5) 51 (10.8) 12 (9.0) Unknown 657 (1.5) 44 (1.4) Health plan, n (%) PPO 32,990 (77.0) 2,347 (75.0) 344 (72.9) 84 (62.7) HMO 6,103 (14.3) 519 (16.6) 87 (18.4) 36 (26.9) CDHP 269 (0.6) 16 (0.5) Other 3,266 (7.6) 233 (7.5) 34 (7.2) 12 (9.0) Unknown 208 (0.5) 14 (0.5) CCI score, mean (SD) 1.2 (1.9) 1.5 (2.2) 1.8 (2.3) 2.3 (2.5) Comorbid condition, n (%) Autoimmune disease 7,745 (18.1) 723 (23.1) 116 (24.6) 53 (39.6) Ulcerative colitis 2,326 (5.4) 238 (7.6) 39 (8.3) 21 (15.7) Crohn’s disease 1,782 (4.2) 175 (5.6) 22 (4.7) 11 (8.2) Type 1 diabetes 1,359 (3.2) 134 (4.3) 18 (3.8) 11 (8.2) Renal insufficiency 5,618 (13.1) 571 (18.3) 105 (22.3) 36 (26.9) Pre-index medications, n (%) Gastric acid-suppressing agents 11,943 (27.9) 1,028 (32.9) 184 (39.0) 51 (38.1) Antibiotics 33,411 (78.0) 2,509 (80.2) 381 (80.7) 103 (76.9) Pre-index healthcare use, n (%) Inpatient admission 13,938 (32.5) 1,307 (41.8) 236 (50.0) 81 (60.5) Outpatient hospital visit 32,584 (76.1) 2,576 (82.3) 404 (85.6) 116 (86.6) ED visit 19,534 (45.6) 1,581 (50.5) 268 (56.8) 77 (57.5) Outpatient office visit 40,064 (93.5) 2,951 (94.3) 455 (96.4) 129 (96.3) For patient privacy reasons and consistent with data reporting practices for the Centers for Medicare and Medicaid Services, data are not shown for cells in which the sample size was 10. Abbreviations. CCI, Charlson comorbidity index; CDHP, consumer-directed health plan; ED, emergency department; HMO, health maintenance organization; PPO, preferred provider organization; SD, standard deviation. 606 P. FEUERSTADT ET AL. 540 patients, the total hospitalization cost of treating a rCDI Discussion episode was 2.2 times that of primary CDI . Our study pro- CDI and rCDI are associated with substantial all-cause HRU vides a broader picture for CDI and rCDI medical costs, in and direct medical costs. Mean total, all-cause, direct medical more current dollars, that are not exclusively related to the costs for patients with rCDI varied by number of rCDI epi- CDI episode/hospitalization. Our study also focuses on a sodes and ranged between $131,000 for patients with 1 younger cohort than most other studies, with included recurrence to more than $200,000 in patients with 3 or more patients between the ages of 18 and 64 years. recurrences in the 12 months following an index episode; After a 6-month baseline period that was devoid of any costs for inpatient stays made up the majority of direct costs CDI claims, the average time from the index CDI episode to for all cohorts. The mean time between rCDI episodes was each rCDI episode was approximately 1 month, similar to stable at about 1 month from the previous episode, with a what has been reported previously . This likely represents standard deviation of approximately 15 d regardless of study the highest risk period for recurrence within the “window of cohort, indicating that the recurrence timing in this cohort vulnerability.” Once the colonic microbiota has been signifi- aged 18–64 years seems predictable. cantly depleted, the antibiotic treatment suppresses the CDI, Prior studies have shown a mean CDI-attributable cost but subsequently there might be insufficient time to restore from $8,911 to $30,049 per patient/admission/episode/ the microbiota to a state where it is protective against reacti- infection (2014 US dollars) . In one single centre study of vation of the latent C. difficile spores. Therefore, rate of recur- rence and time to recurrence could be expected to be similar across episodes, without the presence of other influ- encing factors (e.g. concomitant antibiotics). Because the time to recurrence was relatively brief, we would expect the patient’s health plan at the time of the index CDI to assume most of the costs associated with the subsequent rCDI epi- sode(s). This 1-month period also presents itself as a window of opportunity for interventions that can reduce the risk for future rCDI episodes. The findings of the logistic regression are important particularly because minimally significant fac- Figure 2. The mean (SD) days from one CDI episode to a recurrence (within tors were found that predict multiple recurrences vs. only 1 the 8-week window defining rCDI) was consistent at approximately 1 month and did not depend on the number of rCDI episodes (n¼ 3,735 patients). recurrence. Therefore, it seems that preventing any recur- rence is the key objective, which should be an imperative for all CDI care. Though there are published analyses reporting risk factors for any CDI recurrence, the risk factors for mul- tiple rCDI vs. 1 rCDI have not been extensively studied 8,18–21 before . HRU was high for all patients with an index CDI, with the highest utilization for those with 3 or more recurrences. Other analyses of HRU have shown a similar hospital length of stay of 8.0 d for the index CDI episode, and 9.3 d for a rCDI episode . In 2016, the mean hospital length of stay in the US for all conditions was 4.6 d, indicating that CDI places a higher burden on HRU than the average admission .A single centre study showed similar rates of ICU use to our study, with 9.4% of patients with rCDI having an inpatient admission with ICU stay . An analysis of electronic records from 85 hospitals in the US showed the highest incidence of Figure 3. In the 12-month follow-up period, the mean number of visits per 25 community-onset CDI was in the Northeast and Midwest . patient for the emergency department (ED) or for an inpatient admission was Our results support some of the earlier studies, with the highest for those with 3 or more CDI recurrences, at 4.6 visits and 5.8 visits, respectively. highest rates of rCDI occurring in the Midwest. Table 2. All-cause healthcare resource utilization during 12-month follow-up. No recurrence 1 Recurrence 2 Recurrences 3þ Recurrences Resource utilized (N¼ 42,836) (N¼ 3,129) (N¼ 472) (N¼ 134) Patients with admissions, n (%) 0 Admissions 15,875 (37.1) 898 (28.7) 111 (23.5) 29 (21.6) 1 Admission 14,490 (33.8) 572 (18.3) 68 (14.4) 12 (9.0) 2 Admissions 12,471 (29.1) 1,659 (53.0) 293 (62.1) 93 (69.4) LOS per admission, mean (SD) 7.7 (9.7) 8.3 (7.8) 7.9 (6.2) 8.5 (6.6) Inpatient admission with an ICU Stay, n (%) 2,332 (5.4) 251 (8.0) 27 (5.7) 17 (12.7) Prescriptions per patient, mean (SD) 40.5 (39.5) 51.5 (42.6) 61.6 (46.4) 65.0 (51.0) Abbreviations. ICU, intensive care unit; LOS, length of stay; SD, standard deviation. JOURNAL OF MEDICAL ECONOMICS 607 burden of rCDI, and, therefore, we believe, a strict definition increased the validity of the findings. There may, unfortu- nately, have been cases of recurrent CDI that did not meet our case definition; thus, the reported incidence of rCDI may be an underestimation. The PharMetrics Plus database contains adjudicated claims from payers for a diverse US population. The cost data col- lected reflect the payments made by the health plans or employers, therefore, we believe this data source is suffi- ciently accurate and valid to quantify the economic impact of rCDI to US commercial payers. As the data are solely from the US, they may not be generalizable to other countries. In this study, it is likely that the medical costs for the cohort with 3 or more recurrences are underestimated, because the cost summation stopped at 12 months following the index Figure 4. Total, all-cause, direct medical costs during the 12-month period case of CDI. It is most likely that those patients who suffered after an index CDI episode were substantial for any patient with recurrent CDI, and highest for those with 3 or more recurrences. Inpatient costs made up the 3 or more recurrences incurred additional CDI-related costs majority of the total costs. Costs are adjusted to 2018 dollars. beyond 12 months from the index episode. One limitation of this study is that it did not include indi- Table 3. Multivariate logistic regression for 2þ recurrences vs. 1 recurrence. viduals 65 years, as almost all of these individuals in the Odds ratio 95% CI p Value US are covered by Medicare, and the database used for the Female (vs. male) 1.13 0.93 1.36 .21 analysis contained only commercial claims. Previous research Age, years (vs.18–39) showed that older patients are more susceptible to rCDI, and 40–51 0.99 0.77 1.27 .92 52–58 1.04 0.81 1.34 .76 the per-patient cost of treatment could be higher due to the 59þ 0.78 0.60 1.02 .07 9,26 impact of chronic conditions . Future studies focusing on Geographic region (vs. Midwest) Northeast 1.07 0.84 1.35 .59 the older, frail population would add to the current body of South 0.91 0.72 1.14 .41 evidence. Also, a retrospective analysis of insurance claims West 0.88 0.64 1.21 .42 data is subject to potential miscoding, however, these data Type of benefit plan (vs. PPO) HMO 1.26 0.99 1.59 .06 errors are expected to be random and equal across the HDHP/CDHP 2.17 0.87 5.45 .10 population. There may be missing encounter data when Other 1.07 0.76 1.51 .68 patients paid out of pocket, which would lead to an under- Unknown 1.03 0.29 3.69 .96 CCI Score (vs.0) estimation of costs. To address the key study objective, the 1 1.22 0.92 1.61 .17 study included patients who had a minimum of 18 months 2þ 1.24 0.98 1.55 .07 of continuous enrolment. These inclusion criteria would have Baseline medications Gastric acid suppressing agents (vs. not) 1.14 0.94 1.38 .17 excluded patients who disenroled sooner, including patients Antibiotics (vs. not) 1.00 0.80 1.25 .99 who died due to CDI, rCDI, or other reasons. In addition, Immunosuppressed (vs. not) 1.16 0.95 1.41 .16 individuals were excluded from this study if they lost com- Medical procedures and treatments (vs. none) 1.27 1.04 1.55 .02 Year of index CDI (vs. 2010–2013) mercial health insurance for any reason during the study 2014–2017 1.10 0.92 1.31 .32 period and could not be followed for 12 months; therefore, Inpatient admission in pre-index period (vs. none) 1.23 0.99901.52 .051 patients who became too disabled to work may have been Immunosuppressed defined as taking immunosuppressant agent(s) and/or excluded by this criterion. The impact of this limitation might has autoimmune disease during baseline period. Abbreviations. CCI, Charlson comorbidity index; CDHP, consumer-directed have led to an under-estimation of the resource use and health plan; CDI, Clostridioides difficile infection; CI, confidence interval; HMO, cost findings. The database does not define whether the health maintenance organization; PPO, preferred provider organization. patients had healthcare vs. community-acquired infection. The recurrence rates seen in this study are lower that Because this commercial database does not contain mortality what has been reported elsewhere (between 13.5% and 35% data for patient privacy protection, the analysis cannot dis- 2,26 recurring within 30 d after initial CDI diagnosis) . These cern the reason for disenrolment. As such, the study was lower rates are likely due to our study including a younger designed to focus on depicting economic burden, while (18–64 years) cohort than those who are most frequently unable to report mortality burden of rCDI. 9,26,27 affected by CDI (65 years) . Younger patients tend to have fewer medical comorbidities, less requirement for con- Conclusion comitant antimicrobials, and likely lower rates of CDI recur- rence. The mean CCI values reported here for patients with CDI and rCDI are associated with substantial all-cause HRU any rCDI are slightly lower than what was reported in other and direct medical costs. As expected, repeat recurrences 28–31 observational and randomized, controlled trials in CDI . multiply the economic burden further. Despite a lack of pre- Another potential reason for our lower observed recurrence dictors for multiple vs. 1 rCDI from patient characteristics, the rates was our stringent criteria identifying likely rCDI cases. timing of rCDI recurrence appears predictable. Leveraging The study’s key objective was to quantify the economic this window of opportunity for early interventions that aim 608 P. FEUERSTADT ET AL. [8] Zilberberg MD, Reske K, Olsen M, et al. Risk factors for recurrent at reducing recurrences is key to addressing the economic Clostridium difficile infection (CDI) hospitalization among hospital- burden of rCDI. Future research could consider a prospective ized patients with an initial CDI episode: a retrospective cohort study design of rCDI costs and consider a cohort study. BMC Infect Dis. 2014;14(1):306. aged 65 years. [9] Desai K, Gupta SB, Dubberke ER, et al. Epidemiological and economic burden of Clostridium difficile in the United States: estimates from a modeling approach. BMC Infect Dis. 2016;16(1): Transparency [10] Rodrigues R, Barber GE, Ananthakrishnan AN. A comprehensive Declaration of funding study of costs associated with recurrent Clostridium difficile infec- tion. Infect Control Hosp Epidemiol. 2017;38(2):196–202. The study was funded by Ferring Pharmaceuticals Inc. (Parsippany, NJ). [11] Dubberke ER, Olsen MA. Burden of Clostridium difficile on the healthcare system. Clin Infect Dis. 2012;55(2):S88–S92. [12] van Baal P, Morton A, Meltzer D, et al. Future unrelated medical Declaration of financial/other interests costs need to be considered in cost effectiveness analysis. Eur J Health Econ. 2019;20(1):1–5. Drs. Stong, Nelson, and Dahdal are employees of Ferring [13] Rappange DR, van Baal PHM, van Exel NJA, et al. Unrelated Pharmaceuticals Inc. Dr. Feuerstadt has served as a consultant to and on medical costs in life-years gained: should they be included in eco- the speaker’s bureau for Merck and Co, and has served as a consultant nomic evaluations of healthcare interventions? Pharmacoeconomics. for Ferring Pharmaceuticals Inc. and Roche Pharmaceuticals. Drs. Sacks 2008;10:815–830. and Lang are employees of Precision Health Economic and Outcomes [14] U.S. Bureau of Labor Statistics. Measuring price change in the Research and provided consulting services to Ferring CPI: medical care [Internet]; [cited 2020 Jan 9]. Available from: Pharmaceuticals Inc. https://www.bls.gov/cpi/factsheets/medical-care.htm A peer reviewer on this manuscript has disclosed that they have [15] Nanwa N, Kendzerska T, Krahn M, et al. The economic impact of received consultancy fees from Astellas Pharma. The peer reviewers on Clostridium difficile infection: a systematic review. Am J this manuscript have no other relevant financial relationships or other- Gastroenterol. 2015;110(4):511–519. wise to disclose. [16] Shah DN, Aitken SL, Barragan LF, et al. Economic burden of pri- mary compared with recurrent Clostridium difficile infection in hospitalized patients: a prospective cohort study. J Hosp Infect. Acknowledgements 2016;93(3):286–289. [17] Kelly CP. Can we identify patients at high risk of recurrent Medical writing and editorial support was provided by Agnella Izzo Clostridium difficile infection? Clin Microbiol Infect. 2012;18:21–27. Matic, PhD, CMPP (AIM Biomedical, LLC) and was funded by Ferring [18] Abou Chakra CN, Pepin J, Sirard S, et al. Risk factors for recur- Pharmaceuticals Inc. rence, complications an mortality in Clostridium difficile infection: a systematic review. PLoS One. 2014;9(6):e98400. [19] Escobar GJ, Baker JM, Kipnis P, et al. Prediction of recurrenct Author contributions Clostridium difficile infection using comprehensive electronic med- LS, WN, DND, NS, and KL designed and conducted the study. All authors ical records in an integrated healthcare delivery system. Infect analyzed and interpreted the data, drafted and critically revised the art- Control Hosp Epidemiol. 2017;38(10):1196–1203. icle for important intellectual content, and approved the article for [20] LaBarbera FD, Nikiforov I, Parvathenani A, et al. A prediction publication. model for Clostridium difficile recurrence. J Comm Hosp Intern Med Perspect. 2015;5(1):26033. [21] van Beurden YH, Nezami S, Mulder CJJ, et al. Host factors are more important in predicting recurrent Clostridium difficile infec- Previous presentations tion than ribotype and use of antibiotics. Clin Microbiol Infect. Portions of the data contained in this manuscript appeared in abstract/ 2018;24(1):85.e1–85.e4. poster form at IDWeek 2019, October 2-6, 2019, Poster 2374. This manu- [22] Zhang D, Prabhu VS, Marcella SW. Attributable healthcare script has not been submitted and is not under consideration for publi- resource utilization and costs for patients with primary and recur- cation anywhere else. rent Clostridium difficile infection in the United States. Clin Infect Dis. 2018;66(9):1326–1332. [23] Freeman WJ, Weiss AJ, Heslin KC. Overview of U.S. hospital stays in References 2016: variation by geographic region. HCUP Statistical Brief #246. December 2018. Agency for Healthcare Research and Quality, [1] Ofosu A. Clostridium difficile infection: a review of current and Rockville, MD. www.hcup-us.ahrq.gov/reports/statbriefs/sb246- emerging therapies. Ann Gastroenterol. 2016;29(2):147–154. Geographic-Variation-Hospital-Stays.pdf. [2] Smits WK, Lyras D, Lacy DB, et al. Clostridium difficile infection. [24] Aitken SL, Joseph TB, Shah DN, et al. Healthcare resource utiliza- Nat Rev Dis Primers. 2016;2:16020. tion for recurrence Clostridium difficile infection in a large univer- [3] Centers for Disease Control and Prevention. Clostridioides difficile sity hospital in Houston, Texas. PLoS One. 2014;9(7):e102848. infection [Internet]; [cited 2019 Aug 13]. Available from: https:// [25] Zilberberg MD, Tabak YP, Sievert DM, et al. Using electronic www.cdc.gov/hai/organisms/cdiff/cdiff_infect.html health information to risk-stratify rates of Clostridium difficile [4] Reveles KR, Lee GC, Boyd NK, et al. The rise in Clostridium difficile infection in US hospitals. Infect Control Hosp Epidemiol. 2011; infection incidence among hospitalized adults in the United 32(7):649–655. States: 2001–2010. Am J Infect Control. 2014;42(10):1028–1032. [26] Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile [5] Leffler DA, Lamont JT. Clostridium difficile infection. N Engl J Med. infection in the United States. N Engl J Med. 2015;372(9): 2015;372(16):1539–1548. 825–834. [6] DePestel DD, Aronoff DM. Epidemiology of Clostridium difficile [27] Pechal A, Lin K, Allen S, et al. National age group trends in infection. J Pharm Pr. 2013;26(5):464–475. Clostridium difficile infection incidence and health outcomes [7] Song JH, Kim YS. Recurrent Clostridium difficile infection: risk fac- in United States community hospitals. BMC Infect Dis. 2016;16(1): tors, treatment, and prevention. Gut Liver. 2019;13(1):16–24. 682. JOURNAL OF MEDICAL ECONOMICS 609 [28] Cammarota G, Masucci L, Ianiro G, et al. Randomised clinical trial: [30] Delholm-Lambertsen E, Hall BK, Jørgensen SMD, et al. Cost sav- faecal microbiota transplantation by colonoscopy vs. vancomycin ings following faecal microbiota transplantation for recurrent for the treatment of recurrent Clostridium difficile infection. Clostridium difficile infection. Ther Adv Gastroenterol. 2019;12: Aliment Pharmacol Ther. 2015;41(9):835–843. 1–14. [29] Charlson ME, Pompei P, Ales KL, et al. A new method of classify- [31] McFarland LV, Surawicz CM, Rubin M, et al. Recurrent Clostridium ing prognostic comorbidity in longitudinal studies: development difficile disease: epidemiology and clinical characteristics. Infect and validation. J Chron Dis. 1987;40(5):373–383. Control Hosp Epidemiol. 1999;20(01):43–50.

Journal

Journal of Medical EconomicsTaylor & Francis

Published: Jun 2, 2020

Keywords: Clostridium difficile infection; Clostridioides difficile infection (CDI); recurrent CDI (rCDI); real-world outcomes; healthcare resource utilization; direct medical costs; I10; I19

There are no references for this article.