The economic burden of treatment failure amongst patients with irritable bowel syndrome with constipation or chronic constipation: a retrospective analysis of a Medicaid population

Abstract

Objective:

To compare healthcare resource utilization (HRU) and costs between patients with irritable bowel syndrome with constipation (IBS-C) or chronic constipation (CC) with and without evidence of treatment failure.

Methods:

Claims data from the Missouri Medicaid program were used to identify adults with IBS-C or CC treated for constipation. IBS-C patients were required to have ≥2 constipation therapy claims, and the index date was defined as the date of the first constipation therapy claim within 12 months after an IBS diagnosis. For CC, the index date was defined as the date of the first constipation treatment claim followed by a second claim for constipation treatment or diagnosis between 60 days and 12 months later. Indicators of treatment failure were: switch/addition of constipation therapy, IBS- or constipation-related inpatient/emergency admission, megacolon/fecal impaction, constipation-related surgery/procedure, or aggressive prescription treatments. Annual incremental HRU and costs (public payer perspective) were compared between patients with and without treatment failure. Incidence rate ratios (IRRs) and cost differences are reported.

Results:

In total, 2830 patients with IBS-C and 8745 with CC were selected. Approximately 50% of patients had ≥1 indicator of treatment failure. After adjusting for confounding factors, patients with treatment failure experienced higher HRU, particularly in inpatient days (IRR = 1.75 for IBS-C; IRR = 1.54 for CC) and higher total healthcare costs of $4353 in IBS-C patients and $2978 in CC patients. Medical service costs were the primary driver of the incremental costs associated with treatment failure, making up 71.3% and 67.0% of the total incremental healthcare costs of the IBS-C and CC samples, respectively.

Limitations:

Sample was limited to Medicaid patients in Missouri. Claims data were used to infer treatment failure.

Conclusion:

Treatment failure is frequent among IBS-C and CC patients, and sub-optimal treatment response with available IBS-C and CC therapies may lead to substantial HRU and healthcare costs.

Keywords::

• Irritable bowel syndrome with constipation
• Chronic constipation
• Treatment failure
• Economic burden

Introduction

Irritable bowel syndrome with constipation (IBS-C) and chronic constipation (CC) are functional gastrointestinal disorders defined by persistent symptoms over an extended period of time with an absence of structural or biochemical causes1. IBS is defined by symptoms of abdominal discomfort or pain associated with features of disordered defecation2; IBS-C is one sub-type of IBS that is characterized by the presence of chronic or intermittent abdominal pain and discomfort in association with constipation-predominant bowel habits, including reduced stool frequency, hard and/or lumpy stool consistency, and/or difficult defecation1,2. CC is also defined by persistent symptoms of unsatisfactory defecation, including infrequent bowel movements, hard or lumpy stools, straining, and a sensation of incomplete rectal evacuation, although many patients also experience abdominal symptoms1,3.

Global prevalence of IBS-C is estimated at ∼5–6%, based on a recent meta-analysis of prevalence findings for IBS and a US population-based survey4,5. The prevalence of CC in North America is estimated at 15% based upon the findings of a systematic review and a recent US population-based survey estimated the prevalence of CC at 11.2%6,7. Prevalence estimates for both IBS-C and CC vary depending on the diagnostic criteria used (e.g., Manning, Rome I, or Rome II criteria).

The abdominal and bowel symptoms of IBS-C and CC are chronic, impair patients’ quality-of-life, and may impose a substantial economic burden on patients and payers2,8–12. Since both diseases affect the digestive system and present with symptoms of visceral discomfort and unsatisfactory defecation, the therapeutic approaches for IBS-C and CC are similar. Management options usually involve lifestyle modifications (e.g., exercise and dietary adjustments) and treatment with over-the-counter (OTC) laxatives, bulking agents, and stool softeners3,8,13. Prescription medications approved for the treatment of IBS-C and CC were limited and, at the time this study was conducted, the only prescription product approved by the Food and Drug Administration (FDA) for the treatment of IBS-C (for women only) and CC was lubiprostone. Since the conclusion of this analysis, linaclotide, a new prescription treatment for IBS-C and CC in adults, has been approved by the FDA. Other prescription products commonly used for the treatment of IBS-C and CC include lactulose and polyethylene glycol8,13.

Although past studies have examined the economic impact of IBS and constipation, information specific to IBS-C and CC is scarce. Most studies have only focused on overall IBS14–17, not IBS sub-types, or constipation only rather than chronic constipation18,19. One study specific to IBS-C patients estimated incremental charges of IBS-C, but did not estimate costs18. Two recent studies of IBS-C and CC have analyzed the economic burden specific to these conditions in terms of direct healthcare costs and found annual direct healthcare costs to be $3856 (in 2010 USD) higher for IBS-C patients11 and $3508 (in 2010 USD) higher for CC patients12 compared with age-, gender-, health plan type-, and region-matched controls in a managed care population.

Because there are few prescription treatment options available, OTC medications are the most commonly used treatment for IBS-C and CC13. However, generally they are not reimbursed by insurance payers and are not recorded in administrative claims databases. Therefore, research on treatment patterns and outcomes associated with OTC use in clinical practice has been limited.

An understanding of treatment failure is important to optimize the care of patients with IBS-C or CC and identifying the associated economic consequences of treatment failure informs the need for new therapies and appropriate patient management. To date, no study has investigated the costs of treatment failure associated with either prescription or OTC treatments in the IBS-C or CC patient population. The objective of this study was to compare healthcare resource utilization (HRU) and costs between patients with IBS-C or CC initiating on prescription or OTC medications with and without evidence of treatment failure.

Methods

Data source

The data source for this analysis was the de-identified administrative claims database from Missouri Medicaid. The Missouri Medicaid database (01/01/1997–12/31/2010) contains complete medical and pharmacy dispensing claims for Medicaid-eligible enrollees during the covered years. Medical claims include both professional service and facility claims. Both OTC and prescription claims are available in the pharmacy data. Because the target population is often treated with OTC products, the availability of OTC treatment information in a claims database was an important factor in choosing the Medicaid database for this study.

During the study period, the Missouri Medicaid program covered a comprehensive list of OTC laxatives, bulking agents, and stool softeners for fee-for-service enrollees. Although OTC drugs generally do not require a physician’s prescription they are needed to obtain reimbursement by the Missouri Medicaid program.

Study design

This study used a retrospective cohort design methodology. Patients with IBS-C or CC were identified as two distinct samples and analyzed separately. Sub-group analyses were also conducted for both samples based on whether patients initiated treatment with OTC or prescription treatment on their index date. All sub-group analyses were conducted independently (i.e., no comparison between the OTC and prescription sub-groups). Patients who initiated treatment on a combination of an OTC and a prescription treatment were included in both sub-groups.

IBS-C patients

Patients were identified as having IBS-C if they had at least two independent claims with a diagnosis of constipation (ICD-9-CM code 564.0x) or constipation treatments and one claim with a diagnosis for IBS (ICD-9-CM code 564.1x) in their claims history. The first constipation therapy claim within 12 months after an IBS diagnosis was defined as the index date.

CC patients

Patients were classified as having CC if they had at least two independent claims with a diagnosis of constipation (ICD-9 CM code 564.0x) or constipation treatments that were more than 60 days apart but within 12 months of each other. For the CC sample, the index date was defined as the first constipation therapy followed by a second claim for constipation therapy or diagnosis at least 60 days but less than 12 months later. Patients in the CC sample were also required to have no diagnoses for IBS in the 6-month period preceding the index date.

Sample selection

In both samples, patients were selected if they: (1) were at least 18 years of age on their index date, (2) had continuous Medicaid eligibility for at least 6 months preceding and 12 months following the index date, (3) had no diagnoses for diarrhea (ICD-9 CM code 564.5x) or claims for anti-diarrheal treatments during the 6 months preceding the index date, and (4) had no inpatient stays lasting longer than 30 consecutive days during the 6 months preceding or 12 months following the index date. The baseline period was defined as the 6-month period preceding the index date; the study period was defined as the 12-month period following the index date. The full observation period between patients’ index date and end of continuous eligibility was also considered to estimate the treatment failure rates over a longer period of time.

For each sample (IBS-C and CC) and sub-group (OTC or prescription sub-group), patients were classified into two mutually exclusive cohorts based on whether or not they met the criteria for treatment failure during the study period. Because claims databases do not include clinical indicators of treatment response that can precisely identify treatment failure nor reasons for therapy change, surrogate indicators suggesting clinical treatment failure were developed in conjunction with clinical experts and were used as proxies. Patients with IBS-C or CC often switch if the therapy is not well tolerated or switch or augment therapies when their symptoms are not well controlled8,20. Some physicians may also prescribe more aggressive prescription therapies such as colchicine, misoprostol, or rifaximin if patients cannot find relief with traditional therapies21,22. In some cases, patients may develop even more severe conditions such as a fecal impaction or megacolon23,24. Severe IBS-C or CC can require medical diagnostic tests such as defecography, transit time tests, and radiologic procedures23. In unusual circumstances, colectomy may be required if patients do not respond to pharmacologic therapy, usually as a last resort24,25. Intense abdominal pain associated with IBS-C or CC can lead to emergency department (ED) or inpatient admissions for treatment.

Therefore, in this study, patients with at least one of the following indicators of treatment failure while ‘on treatment’ were included in the ‘treatment failure’ cohort; otherwise, they were included in the ‘without treatment failure’ cohort. The requirement of ‘on treatment’ indicates the event occurred during the use of originally prescribed medication, and was defined as the period during which a patient had a supply of constipation/IBS-C treatment on hand plus an additional 30 days after. Indicators of treatment failure included (1) a switch from one IBS-C/constipation treatment to another or augmentation with an additional IBS-C/constipation treatment; (2) an IBS- or constipation-related (only constipation-related for patients in the CC sample) inpatient stay or ED visit; (3) a diagnosis of megacolon (ICD-9 CM code 564.7x) or fecal impaction (ICD-9 CM code 560.39 or 751.3x); (4) a constipation-related surgery/medical procedure (i.e., colon transit time test, colectomy, biofeedback, anorectal manometry, ileoresection, pulse irrigation, surgical laparoscopy, defogram, Sitzmark test, SmartPill test, or fecal disimpaction); or (5) the use of a more aggressive prescription (i.e., colchicine, misoprostol, or rifaximin).

Outcomes and statistical analyses

Frequency of treatment failure

The frequency of the different indicators of treatment failure was measured during both the 12-month study period and the entire continuous eligibility period following the index date.

Patient characteristics

Patient characteristics, including demographics, comorbid conditions26,27, Charlson comorbidity index (CCI)28, baseline HRU, and costs were summarized and compared. Comparisons between cohorts (patients with and without treatment failure) were made using chi-square tests for discrete variables and Wilcoxon rank-sum tests for continuous variables. For both samples, an exhaustive list of comorbidities was studied, but only those with at least 5% prevalence are reported.

Healthcare resource utilization

HRU end-points measured during the study period included inpatient admissions, inpatient days, ED visits, outpatient visits, long-term care (LTC) days (including LTC stays of less than 30 consecutive days, home health and nursing home services), mental health facility (MHF) days, and home care visits (HCV). Inpatient admissions were defined as the number of times the patient was admitted to an inpatient facility. Inpatient days were defined as the total number of days spent in inpatient facilities. ED visits were defined as the total number of times the patient was admitted to an ED facility, but ED visits resulting in an inpatient stay were counted as inpatient admissions only. Outpatient visits were defined as the total number of visits to an outpatient facility, including office visits. LTC days were defined as the total number of days spent in LTC facilities. MHF days were defined as the total number of days spent in MHFs (inpatient and outpatient). HCVs were defined as the total number of HCVs the patient received.

Healthcare costs

Costs were inflated to 2010 US dollars using the medical care component of the consumer price index; 2010 was selected as it corresponds to the last year covered by the data. The primary end-points were typical medical costs which included inpatient, ED, and outpatient costs. The secondary end-points were non-typical medical costs which included those related to LTC, MHF days, HCV costs, and other medical service costs (e.g., laboratory, radiology, or other ancillary medical services). Total medical service costs were calculated as the sum of typical and non-typical medical costs. Pharmacy costs were also evaluated. Total healthcare costs were calculated as the sum of paid amounts for pharmacy and total medical service costs.

Statistical analyses

Multivariate regression models adjusted for potential confounding factors, including age, gender, ethnicity, CCI, comorbidities with at least 5% prevalence in both cohorts and a statistically significant difference between cohorts, and baseline HRU and costs. HRU was compared between cohorts using incident rate ratios (IRR) adjusting for potential confounding factors with a generalized linear regression model with a log link and a negative binomial distribution. Healthcare costs were compared between cohorts using generalized linear models. P-values and associated 95% confidence intervals (CIs) were calculated using a bootstrapping technique with 500 iterations. All statistical analyses were conducted using SAS 9.2 (SAS Institute, Inc., Cary, NC). A two-sided alpha error of 0.05 was used to declare statistical significance.

Results

Patient characteristics and frequency of treatment failure

A total of 2830 patients with IBS-C and 8745 patients with CC were selected for the analysis (Figure 1). In the CC sample, 6315 patients initiated on an OTC treatment, 2139 patients initiated on a prescription, and 291 patients initiated on both. In the IBS-C sample, 1627 patients initiated on an OTC, 1107 patients initiated on a prescription, and 96 patients initiated on both.

Figure 1. Sample Selection Flowchart.

During the 12-month study period we observed at least one indicator of treatment failure in 46.3% of patients with IBS-C. When the patients’ entire continuous eligibility periods were considered (median observation period length was 4.9 years), we observed at least one indicator of treatment failure in 75.9% of patients (Table 1). In the sub-groups of IBS-C patients initiated on OTC and prescription treatments, the proportions of patients with at least one indicator of treatment failure were 49.4% and 41.6% during the 12-month study period, and were 80.7% and 68.3% when the patients’ entire continuous eligibility periods were considered, respectively.

Table 1. Frequency of treatment failure.

Indicator of treatment failure*	IBS-C				CC
	12 months following the index date		Entire observation period†		12 months following the index date		Entire observation period^†
	% of failures (n = 1311)	% of total (n = 2830)	% of failures (n = 2149)	% of total (n = 2830)	% of failures (n = 4798)	% of total (n = 8745)	% of failures (n = 6894)	% of total (n = 8745)
Any indicator of treatment failure	100.0%	46.3%	100.0%	75.9%	100.0%	54.9%	100.0%	78.7%
Therapy augmentation with another constipation/IBS-C treatment	17.0%	7.9%	11.5%	8.7%	13.1%	7.2%	9.1%	7.2%
Switch from one constipation/IBS-C treatment to another	74.0%	34.3%	90.4%	68.7%	88.4%	48.5%	94.8%	74.6%
IP or ED admission related to constipation or IBS	24.0%	11.1%	27.8%	21.1%	14.0%	7.7%	20.7%	16.3%
Diagnoses indicating treatment failure	3.7%	1.7%	5.9%	4.5%	3.6%	2.0%	8.3%	4.6%
Procedures indicating treatment failure	10.9%	5.1%	14.6%	11.1%	7.0%	3.9%	9.6%	7.6%
Prescription for more aggressive therapies	8.9%	4.1%	13.2%	10.0%	4.5%	2.4%	6.0%	4.7%

IBS-C, irritable bowel syndrome with constipation; CC, chronic constipation; CCI, Charlson comorbidity index; IP, inpatient; ED, emergency department.

*Categories are not mutually exclusive.

†The median observation time was 4.9 years for both IBS-C and CC samples.

Similarly, during the 12-month study period we observed at least one indicator of treatment failure in 54.9% of patients with CC. When the patients’ entire continuous eligibility periods were considered (median observation period length was 4.9 years), we observed at least one indicator of treatment failure in 78.7% of patients. In the sub-group of CC patients initiated on OTC and prescription treatments, the proportions of patients with at least one indicator of treatment failure were 48.5% and 51.8% during the 12-month study period, and were 81.3% and 74.4% when the patients’ entire continuous eligibility periods were considered, respectively.

In both samples, treatment switching was the most frequent indicator of treatment failure. In the IBS-C sample, 34.3% of patients switched therapies during the 12-month period following the index date and 68.7% of patients switched therapies during their entire eligibility period following the index date (Table 1). In the CC sample, 48.5% of patients switched therapies during the 12-month period following the index date and 74.6% of patients switched treatment at some point during their entire eligibility period following the index date. Findings were similar in the OTC/prescription treatment sub-groups in each sample, with switching therapies the most common indicator of treatment failure.

Patient baseline characteristics

Mean ages were 51 years in the IBS-C sample and 57 years in the CC sample. Eighty-seven per cent of the IBS-C sample was female, and 70% of the CC sample was female. In the IBS-C sample, there were significantly fewer Caucasian patients and significantly more African American patients in the treatment failure cohort (Table 2). However, no significant differences in race distribution were observed in the CC sample. The prevalence of comorbidities was similar between cohorts at baseline. There was a high prevalence of chronic conditions in both samples; for example, in both the IBS-C and CC samples, we observed a prevalence of chronic pulmonary disease and hypertension of over 45%. Furthermore, in the IBS-C sample we observed a prevalence of over 40% in depression and psychoses. In the IBS-C sample, the only comorbidity with a significant difference between cohorts was drug abuse. In the CC sample, the prevalence rates of diabetes with complications, drug abuse, renal failure, and alcohol abuse were all significantly different between cohorts.

Table 2. Patient baseline characteristics.

	IBS-C (n = 2830)			CC (n = 8745)
	With treatment failure	Without treatment failure	p Value	With treatment failure	Without treatment failure	p Value
Number of Patients	1311	1519		4798	3947
Demographics
Age at treatment index date, mean ± SD	51.3 ± 16.2	51.1 ± 16.3	0.715	57.3 ± 18.0	57.1 ± 18.3	0.797
Female, n (%)	1124 (85.7)	1329 (87.5)	0.171	3299 (68.8)	2786 (70.6)	0.065
Ethnicity, n (%)
Caucasian	1062 (81.0)	1299 (85.5)	0.001*	3528 (73.5)	2869 (72.7)	0.376
African American	192 (14.6)	172 (11.3)	0.008*	1096 (22.8)	930 (23.6)	0.428
Asian	0 (0.0)	1 (0.1)	0.353	9 (0.2)	2 (0.1)	0.072
Hispanic	3 (0.2)	4 (0.3)	0.854	21 (0.4)	22 (0.6)	0.426
Other	54 (4.1)	43 (2.8)	0.060	144 (3.0)	124 (3.1)	0.705
CCI, mean ± SD	2.3 ± 2.3	2.3 ± 2.4	0.984	2.6 ± 2.6	2.4 ± 2.6	0.001*
Presence of other physical and mental comorbid conditions†, n (%)
Chronic pulmonary disease	729 (55.6)	791 (52.1)	0.060	2226 (46.4)	1780 (45.1)	0.226
Hypertension (uncomplicated and complicated)	669 (51.0)	777 (51.2)	0.948	2745 (57.2)	2204 (55.8)	0.198
Depression	639 (48.7)	736 (48.5)	0.878	1383 (28.8)	1080 (27.4)	0.130
Psychoses	587 (44.8)	662 (43.6)	0.524	1442 (30.1)	1144 (29.0)	0.275
Fluid and electrolyte disorders	360 (27.5)	396 (26.1)	0.405	988 (20.6)	796 (20.2)	0.624
Diabetes, complicated	116 (8.8)	137 (9.0)	0.874	721 (15.0)	513 (13.0)	0.007*
Diabetes, uncomplicated	294 (22.4)	361 (23.8)	0.399	1429 (29.8)	1088 (27.6)	0.023
Deficiency anemias	272 (20.7)	300 (19.7)	0.510	858 (17.9)	692 (17.5)	0.670
Rheumatoid arthritis/collagen vascular diseases	252 (19.2)	312 (20.5)	0.382	641 (13.4)	539 (13.7)	0.687
Misc. neurological disorders	262 (20.0)	274 (18.0)	0.188	856 (17.8)	714 (18.1)	0.763
Hypothyroidism	235 (17.9)	259 (17.1)	0.541	572 (11.9)	469 (11.9)	0.955
Peptic ulcer disease excluding bleeding	218 (16.6)	240 (15.8)	0.551	407 (8.5)	329 (8.3)	0.805
Valvular disease	210 (16.0)	231 (15.2)	0.553	763 (15.9)	568 (14.4)	0.050
Peripheral vascular disorders	217 (16.6)	213 (14.0)	0.062	875 (18.2)	689 (17.5)	0.343
Obesity	180 (13.7)	238 (15.7)	0.147	508 (10.6)	421 (10.7)	0.906
Congestive heart failure	187 (14.3)	193 (12.7)	0.225	937 (19.5)	766 (19.4)	0.886
Drug abuse	165 (12.6)	154 (10.1)	0.040*	371 (7.7)	253 (6.4)	0.017*
Solid tumor without metastasis	129 (9.8)	177 (11.7)	0.122	535 (11.2)	414 (10.5)	0.322
Alcohol abuse	78 (5.9)	98 (6.5)	0.581	362 (7.5)	249 (6.3)	0.024*
Renal failure	52 (4.0)	64 (4.2)	0.741	353 (7.4)	217 (5.5)	<0.001*
Baseline Healthcare resource utilization, Mean ± SD
IP admissions	0.4 ± 0.9	0.4 ± 0.8	0.163	0.5 ± 1.0	0.4 ± 0.9	<0.001*
IP days	2.1 ± 5.3	1.7 ± 4.6	0.126	4.3 ± 9.9	3.5 ± 9.1	<0.001*
ED admissions	1.7 ± 3.1	1.4 ± 2.6	0.015*	1.0 ± 1.9	0.8 ± 1.7	<0.001*
OP visits	13.5 ± 8.9	12.2 ± 8.2	<0.001*	9.9 ± 8.1	9.4 ± 7.9	0.0003*
Baseline Healthcare Costs, Mean ± SD	8616 ± 8059	7452 ± 7620	<0.001*	7971 ± 10,795	7031 ± 8169	<0.001*
Medical costs	5235 ± 6137	4405 ± 6094	<0.001*	5112 ± 7343	4521 ± 6901	<0.001*
Pharmacy costs	3381 ± 4201	3047 ± 3415	0.020*	2859 ± 7026	2509 ± 3279	0.003*

IBS-C, irritable bowel syndrome with constipation; CC, chronic constipation; IP, inpatient; ED, emergency department; OP, outpatient.

*p Value <0.05.

†Only comorbid conditions with at least 5% prevalence in both groups are shown.

During the 6-month baseline period, in both samples, patients in the treatment failure cohort had significantly higher healthcare costs. In the IBS-C sample, patients with treatment failure also had significantly more ED admissions (1.7 vs 1.4; p = 0.015) and outpatient visits (13.5 vs 12.2; p < 0.001) than patients without treatment failure. Similarly, in the CC sample, treatment failure was associated with significantly higher ED admissions (1.0 vs 0.8; p < 0.001), outpatient visits (9.9 vs 9.4; p < 0.001), and inpatient admissions (0.5 vs 0.4; p < 0.001).

Healthcare resource utilization and costs

During the 12-month study period, patients with treatment failure had significantly higher rates of all HRU compared to patients without treatment failure (Table 3). In both samples, patients with treatment failure also had significantly higher total costs compared to patients without treatment failure. In the IBS-C sample, among patients with treatment failure, mean total healthcare costs were $18,886 vs $13,897 among patients without treatment failure. The total incremental cost was $4353 after adjusting for potential confounding factors (Table 4). In the CC sample, mean total healthcare costs were $18,617 and $14,929 among patients with and without treatment failure, respectively. The total incremental cost was $2978 after adjusting for potential confounding factors.

Table 3. Comparison of healthcare resource utilization between patients with and without treatment failure.

	IBS-C			CC
	Incidence per patient year		Adjusted IRR† IRR (95% CI)	Incidence per patient year		Adjusted IRR† IRR (95% CI)
	With treatment failure	Without treatment failure		With treatment failure	Without treatment failure
All patients
IP admissions	0.92	0.52	1.66 (1.51–1.81)*	0.98	0.62	1.52 (1.44–1.59)*
IP days	4.77	2.53	1.75 (1.68–1.82)*	5.76	3.58	1.54 (1.51–1.57)*
OP visits	28.07	23.13	1.16 (1.14–1.17)*	23.00	19.91	1.12 (1.11–1.13)*
ED visits	3.42	2.6	1.27 (1.22–1.33)*	2.34	1.7	1.35 (1.31–1.39)*
HCV	13.23	10.56	1.23 (1.20–1.26)*	17.83	16.18	1.09 (1.08–1.10)*
LTC days	1.12	0.68	1.67 (1.54–1.81)*	1.54	0.73	2.05 (2.05–2.05)*
MHF days	11.04	8.71	1.16 (1.13–1.19)*	8.64	8.17	1.00 (0.99–1.02)
OTC treatment sub-group
IP admissions	0.95	0.55	1.61 (1.43–1.80)*	1.02	0.6	1.65 (1.56–1.74)*
IP days	4.86	2.67	1.72 (1.63–1.81)*	6.02	3.53	1.67 (1.63–1.71)*
OP visits	28.07	23.46	1.14 (1.12–1.16)*	23.12	19.57	1.13 (1.12–1.14)*
ED visits	3.53	2.62	1.28 (1.21–1.35)*	2.48	1.71	1.39 (1.34–1.44)*
HCV	14.44	11.08	1.24 (1.21–1.28)*	18.15	16.16	1.09 (1.08–1.11)*
LTC days	0.9	0.94	NA	1.59	0.9	NA
MHF days	11.97	9.21	1.25 (1.22–1.29)*	9.26	8.37	1.05 (1.04–1.07)*
Prescription treatment Sub-group
IP admissions	0.87	0.47	1.80 (1.56–2.09)*	1.02	0.63	1.52 (1.38–1.66)*
IP days	4.61	2.33	1.87 (1.75–2.00)*	6.1	3.47	1.64 (1.58–1.71)*
OP visits	28.14	22.83	1.17 (1.15–1.20)*	23.66	20.88	1.10 (1.08–1.12)*
ED visits	3.2	2.7	1.16 (1.08–1.24)*	2.29	1.51	1.53 (1.44–1.62)*
HCV	11.71	9.87	1.28 (1.23–1.32)*	18.02	16.05	1.09 (1.07–1.12)*
LTC days	1.45	0.35	NA	1.49	0.24	NA
MHF days	8.98	8.15	1.04 (1.00–1.09)*	7.38	6.66	1.20 (1.16–1.23)*

IBS-C, irritable bowel syndrome with constipation; CC, chronic constipation; IP, inpatient; OP, outpatient; ED, emergency department; HCV, home care visits; LTC, long-term care; MHF, mental health facility.

*p Value <0.05.

†An IRR >1 indicates that patients with treatment failure had a higher incidence rate of that healthcare resource during the study period. Multivariate regression models adjusted for potential confounding factors, including age, gender, ethnicity, CCI, comorbidities with at least 5% prevalence in both groups and a statistically significant difference, and baseline HRU.

Table 4. Comparison of healthcare costs between patients with and without treatment failure.

	IBS-C			CC
	Patients with treatment failure	Patients without treatment failure	Adjusted^† incremental costs associated with treatment failure	Patients with treatment failure,	Patients without treatment failure,	Adjusted incremental costs associated with treatment failure†
	Mean ± SD	Mean ± SD		Mean ± SD	Mean ± SD
	[A]	[B]	∼[A] − [B]	[A]	[B]	∼[A] − [B]
All patients
Total medical costs ($)	11,021 ± 11,447	7497 ± 8994	3106*	11,503 ± 14,667	9000 ± 12,326	1996*
Typical medical costs	6431 ± 8685	4138 ± 6071	2091*	5756 ± 11,029	4010 ± 8895	1464*
IP costs	2179 ± 5930	1007 ± 3772	1080*	2575 ± 8892	1534 ± 7255	843*
OP costs	3352 ± 3821	2547 ± 3311	706*	2635 ± 3708	2057 ± 3132	510*
ED costs	900 ± 2167	584 ± 1641	306*	545 ± 1462	419 ± 1312	112*
Other medical costs	4590 ± 7659	3359 ± 6215	1014*	5747 ± 8925	4990 ± 8316	532*
HCV costs	2911 ± 6233	2135 ± 5163	684*	4151 ± 7927	3661 ± 7045	370*
LTC costs	46 ± 285	23 ± 247	21*	71 ± 488	29 ± 282	37*
MHF costs	1107 ± 4156	793 ± 2618	251	790 ± 3030	806 ± 3587	−56
Other medical service costs	525 ± 1466	409 ± 1464	58	735 ± 2460	494 ± 1708	180*
Pharmacy costs ($)	7864 ± 8972	6400 ± 7655	1247*	7114 ± 15,040	5929 ± 7664	982*
Constipation/IBS-C treatment costs	171 ± 285	105 ± 274	66*	127 ± 246	86 ± 147	43*
Prescription	97 ± 267	71 ± 272	28*	56 ± 183	42 ± 145	17*
OTC	74 ± 97	34 ± 63	38*	72 ± 168	44 ± 64	27*
Other drug costs	7693 ± 8900	6295 ± 7596	1181*	6986 ± 15,016	5844 ± 7642	939*
Total costs ($)	18,886 ± 15,835	13,897 ± 13,479	4353*	18,617 ± 22,837	14,929 ± 15,861	2978*
OTC treatment sub-group
Total medical costs ($)	11,590 ± 11,752	7711 ± 9088	3415*	12,102 ± 15,375	9159 ± 12,610	2409*
Typical medical costs	6688 ± 9270	4203 ± 6276	2245*	6091 ± 11,651	4022 ± 9165	1745*
IP costs	2313 ± 6280	1122 ± 4193	1085*	2801 ± 9363	1640 ± 7739	977*
OP costs	3397 ± 3921	2519 ± 3378	779*	2706 ± 4002	1957 ± 2896	634*
ED costs	979 ± 2390	562 ± 1372	380*	584 ± 1510	425 ± 1355	134*
Other medical costs	4901 ± 7267	3509 ± 6251	1171*	6012 ± 9385	5138 ± 8376	663*
Pharmacy Costs ($)	8223 ± 8685	6257 ± 6467	1692*	7037 ± 9953	6006 ± 7997	800*
Constipation/IBS-C treatment costs	155 ± 220	67 ± 112	85*	120 ± 245	69 ± 131	51*
Other drug costs	8068 ± 8638	6191 ± 6450	1608*	6917 ± 9913	5937 ± 7986	750*
Total costs ($)	19,813 ± 15,890	13,969 ± 12,697	5108*	19,139 ± 20,211	15,165 ± 16,268	3209*
Prescription treatment sub-group
Total Medical costs ($)	10,013 ± 10,747	7419 ± 8932	2240*	10,974 ± 13,959	8074 ± 10,045	2383*
Typical medical costs	5892 ± 7378	4163 ± 5773	1545*	5626 ± 10,466	3684 ± 6827	1677*
IP costs	1905 ± 5155	866 ± 3082	943*	2380 ± 8259	1092 ± 5161	1066*
OP costs	3237 ± 3581	2669 ± 3237	468*	2723 ± 3473	2252 ± 2973	421*
ED costs	749 ± 1624	628 ± 1949	135	523 ± 1420	340 ± 984	190*
Other medical costs	4121 ± 8187	3255 ± 6339	695	5348 ± 8124	4390 ± 7399	706*
Pharmacy costs ($)	7294 ± 9505	6731 ± 8908	552	7370 ± 24,068	6127 ± 7951	929
Constipation/IBS-C treatment costs	202 ± 366	157 ± 388	43*	154 ± 341	156 ± 295	6
Other drug costs	7091 ± 9388	6573 ± 8815	509	7216 ± 24,050	5971 ± 7895	923
Total costs ($)	17,306 ± 15,690	14,150 ± 14,556	2793*	18,344 ± 29,626	14,201 ± 14,474	3312*

IBS-C, irritable bowel syndrome with constipation; CC, chronic constipation; IP, inpatient; OP, outpatient; ED, emergency department; HCV, home care visits; LTC, long-term care; MHF, mental health facility.

*p Value <0.05.

†Multivariate regression models adjusted for potential confounding factors, including age, gender, ethnicity, CCI, comorbidities with at least 5% prevalence in both groups and a statistically significant difference, and baseline costs.

For both samples, medical service costs made up ∼60% of the total healthcare costs. Average annual medical service costs for IBS-C patients were $11,021 and $7497 for the treatment failure and without treatment failure cohorts, respectively. After adjusting for potential confounding factors, patients in the treatment failure cohort had $3106 higher medical service costs than patients without treatment failure (p < 0.001). Medical service costs were the primary cost driver of the incremental costs associated with treatment failure, making up 71.3% of the total incremental healthcare costs. In the CC sample, annual medical service costs were $11,503 and $9000 for the treatment failure and without treatment failure cohorts, respectively. After adjusting for potential confounding factors, the incremental medical service cost was $1996 (p < 0.001). Similar to the IBS-C sample, the incremental medical service costs made up 67.0% of the total incremental healthcare costs. The individual cost components observed to be highest between cohorts were pharmacy (29% for IBS-C; 33% for CC) and inpatient (25% for IBS-C; 28% for CC) costs. Constipation/IBS-C treatment costs contributed very little to the total incremental cost between cohorts (2% for IBS-C; 1% for CC) (Table 4). Consistent results were found when OTC and prescription medication sub-groups were analyzed, for both the IBS-C and CC samples.

Discussion

To the best of our knowledge, this study is the first published report to evaluate the rate and estimate the costs of treatment failure in IBS-C and CC patients who received physician-prescribed medication, including OTC medication using data from an administrative claims database. The study compared various measures of HRU and costs between cohorts of patients with and without treatment failure. A high rate of treatment failure was observed and significantly higher incidence of medical service resource use and higher total healthcare costs were observed in IBS-C and CC patients who experience treatment failure compared to those who do not experience treatment failure. This suggests a substantial unmet need for effective therapies for both IBS-C and CC patients and the need for appropriate initial treatment, which could limit HRU and costs associated with sub-optimal treatment response. These findings also suggest that there may be higher overall healthcare costs incurred when implementing formulary step edits (i.e., requiring patients to fail therapy before other treatments are available).

The findings of this study show a notable unmet need for effective therapies, given the high rate of treatment failure over a 12-month period and over the entire follow-up period. The study estimated that 46.3% of patients with IBS-C and 54.9% of patients with CC incurred at least one of the observable indicators of treatment failure while on their original prescribed medications (either prescription or OTC) during the 12 month study period. This rate, however, could be an under-estimation. In this study, in addition to medication switching and augmentation, patients may have experienced four additional types of indicators of treatment failure. To ensure the treatment failure events were attributable to the original prescribed constipation medication, we required the event to occur during the course of treatment, which is a very conservative means of defining treatment failure. Given such a conservative method, indicators of treatment failure occurring after treatment discontinuation were not included.

The reasons for the higher costs associated with treatment failure in this study are unknown, and IBS-C- or CC-specific HRU may account for a portion of the incremental costs. Treatment failure could exacerbate IBS-C- and CC-related comorbidities and complications, which could account for a substantial portion of the costs. Additionally, patients with unresolved gastrointestinal symptoms may exhibit increased healthcare seeking and may receive increased medical attention for general comorbid conditions as well as IBS-C or CC. Overall, patients experiencing treatment failure (i.e., those not being treated effectively) used more medical resources and incurred higher costs because their condition was not adequately managed. There is a need to understand whether early treatment with effective therapy options will result in reduced HRU and costs. The high economic burden associated with IBS-C and CC may be alleviated or reduced if IBS-C and CC patients’ symptoms can be better controlled with more effective therapies.

Limitations

Study limitations include the fact that the HRU and costs data reported may not be causally related to unmet needs and treatment failure among patients with IBS-C or CC. The study can, therefore, establish only a temporal association between these factors. In addition, the population studied was limited to patients living in Missouri and enrolled in Medicaid. These patients are likely to be of low socioeconomic status and have a greater comorbidity burden than privately-insured patients. Thus, it may not be possible to generalize our conclusions to the US population as a whole. Also, although the contributions of comorbid disease states are important with regard to HRU and costs, particularly given the high prevalence of these in the study population, their impact was not disaggregated in this analysis. For example, HRU and costs attributable to comorbid conditions, as stratified by patients’ IBS-C- or CC-related treatment failure status, were not evaluated or reported in this study. Further, the possibility that treatment failure in patients with IBS-C or CC might influence clinical activity of seemingly unrelated comorbid conditions or healthcare provider decisions to use diagnostic and treatment services in the management of comorbid conditions was not evaluated in the study.

Another limitation of this study is that treatment failure cannot be precisely identified in claims data; claims data were used to infer treatment failure based on the presence of different potential indicators of treatment failure. For example, treatment switching was used as an indicator of treatment failure; however, it is possible that in some cases, the switches were not due to medical reasons. In addition, because the definition of treatment failure used in the investigation includes the requirement for additional tests, procedures, treatments, and medical services, patients with treatment failure will likely have higher HRU and costs. Furthermore, because the diagnostic criteria for both of these disorders are symptom-based and the criteria used to differentiate IBS-C and CC (e.g., Rome criteria) are not routinely used in clinical practice1, these two functional gastrointestinal disorders may not be robustly separated in real world health data.

Finally, the study is subject to inherent limitations of retrospective studies derived from administrative claims databases. Claims data may be subject to inaccuracies in coding diagnoses and procedures and do not include complete information on disease severity. Nevertheless, these limitations should affect both cohorts equally and, thus, not significantly impact the interpretation of our findings.

Conclusion

Treatment failure is frequent among IBS-C and CC patients, and is associated with a significant burden in terms of both HRU and healthcare costs.

Transparency

Declaration of funding

This study was sponsored by Forest Research Institute, and Ironwood Pharmaceuticals, Inc.

Declaration of financial/other relationships

A. Guerin, M. S. Kaminsky, and E. Q. Wu are employees of Analysis Group Inc. which has received consultancy fees from Forest Research Institute. R. T. Carson is an employee of Forest Laboratories, Inc and owns stock/stock options. B. Lewis was an employee of Ironwood Pharmaceuticals Inc. at the time the study was conducted. D. Yin was an employee of Forest Research Institute at the time the study was conducted. CMRO Peer Reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors thank Douglas C. A. Taylor for his review and contributions to the writing of the manuscript and Steven J. Shiff, MD for providing clinical consultation on patient identification criteria.