Resource use and societal costs for Crohn's disease in Sweden

ORIGINAL ARTICLE

Resource Use and Societal Costs for Crohn’s Disease in Sweden Johan Mesterton, MSc,* Linus Jo¨nsson, PhD,* Sven H.C. Almer, MD,† Ragnar Befrits, MD,‡ Ingalill Friis-Liby, MD,§ and Stefan Lindgren, MDk

Background: The usual onset of Crohn’s disease (CD) is

treated or do not respond to treatment. Thus, total costs of care might be reduced by efficient treatment.

between 15 and 30 years of age, thus affecting people during their most economically productive period in life.

(Inflamm Bowel Dis 2009;15:1882–1890)

Methods: This study intended to estimate societal costs and health-related quality of life (HRQoL) in Swedish patients in different stages of CD. Cross-sectional data on disease activity (measured with the Harvey–Bradshaw Index [HBI]), direct medical resource use, work productivity, and HRQoL (assessed using the 15D instrument) were collected for 420 patients by questionnaires to patients, to the treating physician, and from medical records. Based on HBI, current treatment, and response to treatment, patients were classified into the following disease states: Remission, Response, Active, Refractory, and Surgery. Results: The average 4-week cost per patient in 2007 was estimated at €721 (USD 988), of which 64% was due to lost productivity. The total 4-week cost of care was €255 (USD 349) in Remission, €831 (USD 1138) in Response, €891 (USD 1220) in Active, €1360 (USD 1864) in Refractory, and €16984 (USD 23269) in Surgery. HBI was the most important predictor of costs of care—a 1-point increase in HBI increased total costs by 25% (P < 0.001). HRQoL differed between the disease states: 0.92 in Remission, 0.90 in Response, 0.82 in Active, 0.81 in Refractory, and 0.77 in Surgery. Conclusions: Patients in remission have the lowest costs and the highest HRQoL. Patients responding to treatment have lower costs of care than patients with high disease activity who are not Received for publication February 9, 2009; Accepted February 23, 2009. From the *i3 Innovus, Stockholm, Sweden; †Gastroenterology and Hepatology/IKE, Linko¨ping University, Linko¨ping, Sweden, ‡Gastroenterology Division, Karolinska University Hospital, Stockholm, Sweden, § Gastroenterology/Hepatology Division, Sahlgrenska University Hospital, Gothenburg, Sweden, kGastroenterology/Hepatology Division, University Hospital MAS, Malmo¨, Sweden. Supported by UCB Nordic A/S. The funding source played no role in the design, methods, data collection, analysis, or interpretation of the results. Publication of this study was not contingent on the sponsor’s approval. Reprints: Stefan Lindgren, MD, PhD, Gastroenterology-Hepatology Division, Department of Clinical Sciences, University Hospital MAS, S-205 02 Malmo¨, Sweden (e-mail: [email protected]). C 2009 Crohn’s & Colitis Foundation of America, Inc. Copyright V DOI 10.1002/ibd.20939 Published online 30 April 2009 in Wiley InterScience (www.interscience. wiley.com).

1882

Key Words: Crohn’s disease, costs, quality of life, HarveyBradshaw index

n Sweden 20,000 people suffer from Crohn’s disease (CD).1,2 In the US and Canada, the number of CD patients has been estimated at 630,000, and in Europe between 23,000 and 41,000 new cases of CD are diagnosed annually.3 The usual onset of this chronic, episodic inflammatory bowel disease (IBD) is between 15 and 30 years of age, thus affecting people during their economically most productive period in life. Since life-expectancy is almost normal, patients with CD will live most of their lifetime with their IBD. Most CD patients will require long-term antiinflammatory treatment, sometimes more intensive. Periodically, some patients will require inpatient care and the majority of them will still undergo surgical interventions at some stages during the course of their disease. CD therefore causes considerable medical costs. Moreover, active CD is associated with significant reductions in patients’ ability to work,4 entailing considerable costs for society. With the introduction of novel biological treatments for CD and the increasing demand from healthcare providers to consider the economic impact of different therapeutic alternatives, there is a need to estimate the costs of CD and to increase the understanding of the societal burden of the disease. Although a few studies published during the last decade have analyzed the costs of CD,5–9 many previous estimates do not allow for an adequate stratification of the patients based on disease severity. The primary objective of this study was therefore to estimate the societal costs, both indirect costs and direct medical costs, at different stages of CD in Sweden. Secondary objectives were to identify predictors of costs of care and to estimate health-related quality of life (HRQoL) in relation to disease severity.

I

MATERIALS AND METHODS Study Design In this noninterventional, cross-sectional study, 420 patients were enrolled between August 2007 and February Inflamm Bowel Dis Volume 15, Number 12, December 2009

Inflamm Bowel Dis Volume 15, Number 12, December 2009

2008 at 4 study centers; Karolinska University Hospital, Malmo¨ University Hospital, the University Hospital in Linko¨ping, and Sahlgrenska University Hospital. The study centers were chosen on the basis of having a large, representative, and nonselected catchment area. No patients from private hospitals or private outpatient clinics were included in the study. The study was approved by the regional ethics committee in Lund. A list of all patients with a clinical diagnosis of CD was compiled at each participating center. A random sample of patients was drawn from each list, with an initial overrecruitment compared to the number of patients aimed at in order to counter possible problems due to patients not fulfilling inclusion criteria or inability to contact patients. To be considered for enrolment into the study patients were screened to ensure that they fulfilled all the following inclusion criteria: diagnosed with CD at least 12 months previous to study start; at least 18 years old; understanding written and spoken Swedish or English sufficiently well to be able to complete study procedures; and willing to participate in the study and signed the patient Informed Consent Form before any study-related procedure. Exclusion criteria from study participation were: physical or mental health problems interfering with the ability to participate in study procedures; current participation in a clinical study with blinded treatment.

Data Collection Data were collected on patient and disease characteristics, treatment history, resource use, and HRQoL through a questionnaire completed by the patient, a research nurse, and the treating physician, as well as a questionnaire completed through an interview with the patient during a personal visit (study visit).

Disease Activity Disease activity was assessed using the Harvey–Bradshaw Index (HBI).10 To achieve a more stable estimate of the number of liquid stools per day, the index was slightly modified by asking patients to state the number of liquid stools during the last 7 days. This number was then divided by 7 to obtain the number of liquid stools per day, resulting in a continuous index.

Resource Use Data on patient resource utilization were collected retrospectively from the patient’s records and through the interview with the patient. The use of inpatient care, medical treatments, outpatient care, diagnostic tests, complementary medicine, and lost productivity were assessed. The recall period was different for these categories, depending on the expected frequency of the resource use.

Societal Costs for CD in Sweden

Hospitalizations For all hospitalizations during the last 12 months, the treating physician noted the admission date and discharge date of the hospitalization and the ward (surgical, medicine, geriatric, intensive care unit, or other). Furthermore, it was recorded if the patient underwent surgery (strictureplasty, resection, or other) during the hospitalization.

Medical Treatments Data on the use of pharmaceuticals deemed relevant for the treatment of CD were collected. The following medical treatments were recorded for each patient: antibiotics, peroral and rectal 5-ASA, peroral and rectal corticosteroids, immunosuppressive drugs (azathioprine, methotrexate and mercaptopurine), and anti-TNF therapy (infliximab and adalimumab). For all ongoing treatments, the start date of the treatment, the current dosage, and the patient’s response to the treatment were recorded. For all treatments ended during the last 3 years, data were collected on when treatment was discontinued, the dosage, and the reason for discontinuation.

Outpatient Care The number of outpatient emergency visits and visits to a general practitioner, specialist physician, nurse, occupational therapist, ergonomist, or psychologist during the preceding 3 months were recorded.

Diagnostics/Imaging The number of endoscopies, abdominal x-rays, computed tomography (CT)-scans, and magnetic resonance imagings (MRIs) performed during the last 12 months preceding the visit were recorded.

Complementary Medicine Patients were asked to state whether they had used any complementary medicine during the last 3 months. The patients stated the amount spent on acupuncture, homeopathy, and other complementary medicine.

Lost Productivity Lost productivity was estimated using the Work Productivity and Activity Impairment Questionnaire (WPAI),11 which has been used and validated in CD.12 Patients who were employed at the time of the study visit were asked to consider the impact of their disease on their ability to work over the previous 7 days before the study visit. The patients were asked both to state the number of days they had been absent from work because of their disease (absenteeism) and to estimate the reduction in productivity while at work due to the disease (presenteeism). The percentage of overall impairment of the patient’s work capacity due to CD was obtained by summing these 2 components.

1883

Mesterton et al

Cost Calculations Total costs were calculated by multiplying the resources used by the corresponding unit costs. Costs were calculated over a 4-week period, since this was deemed a relevant time period to observe clinical changes in disease activity. Four weeks has also previously been used as a timeframe for estimating costs of CD.7 For surgical interventions, the cost was only included if the patient had been discharged from the operation-related hospital admission within 6 weeks prior to the study visit. Hence, only for patients classified to the surgery states were the costs of surgery included in the total cost. The reason for this is that patients can respond quickly to surgical treatment and the only way to fairly relate the cost of surgery to current disease activity was to capture only recently undergone operations.13 For the same reason, only pharmaceuticals that the patient was treated with at the time of the visit were included in the cost calculations. Unit costs for the resource use items collected were primarily collected from regional price lists of the 4 participating university hospitals.14–17 Hospitalizations were priced using per diem costs and costs of surgical procedures were derived using the hospital treatment classification system DRG (diagnosis-related group). The average cost was calculated using DRG-codes 148 and 149 (major bowel procedures – with and without complications). All pharmaceuticals were priced using the largest available pack size and strength.18 The value of lost productivity was estimated using the human capital approach.19 According to this approach the cost of a person’s reduced productivity is the amount an employer would be willing to pay for that production, which was assumed to be the average gross weekly salary20 plus employer contributions21 in Sweden. No cost of lost leisure time was included. Unit costs that were not available for 2007 were inflated using the Swedish consumer price index.22 Unit costs were transformed into Euros using the average 2007 exchange rate; €1 ¼ 9.25 SEK (€1 ¼ USD 1.37).23 In Table 1 the unit costs for all resources are presented.

Quality of Life HRQoL was assessed using the patient-administered instrument 15D,27 which includes 15 domains representing different aspects of importance for quality of life. An additive algorithm is used for deriving a total utility score between 0 and 1, representing the overall health utility of the patient. A utility of 0 is assumed to be equivalent to being dead, while a utility of 1 corresponds to having perfect health. Health utilities can be used together with life expectancy to derive quality-adjusted life years (QALYs). QALY is a valuable tool for evaluating the impact of different therapeutical alternatives, as it can capture effects both on life expectancy and on quality of life.19

1884


TABLE 1. Unit Costs, 2007 € (2007 Exchange Rate €1¼USD 1.37)23 Type of Resource Tests and diagnostic imaging Endoscopy Bowel x-ray CT scan MRI Surgery Major bowel operation (DRGs 148/149) Hospital stay, per day Surgical ward Medical ward Geriatric ward Intensive care unit Outpatient visits Emergency visit GP Specialist physician (gastroenterologist) Nurse Occupational therapist Ergonomic therapist Psychologist Drugs Peroral 5-ASAs (per g) Rectal 5-ASAs (per dose) Peroral corticosteroids Prednisolone (per g) Budesonide (per g) Rectal corticosteroids Prednisolone (per dose) Budesonide (per dose) Immunosuppressive drugs Azathioprine (per g) Methotrexate (per g) Mercaptopurine (per g) Anti-TNF therapy Infliximab (per g) Adalimumab (per g) Antibiotics Metronidazole (per g) Productivity costs Average wage per week (gross, incl. employer’s contributions)

Unit Cost (€)

References

433 320 268 493

14 14 14 14

13,181

14-17

539 467 444 1337

14,15,17 14,15 14,15,17 41

236 91 193 77 97 86 90

14 14 14 14 14 14,42 43

0.75 2.4

18 18

11 433

18 18

3.4 6.4

18 18

3.0 61 11

18 18 18

5898 15,053

18 18

0.68

18

940

20,21

Classication of Patients into Disease States To investigate how resource use and quality of life varies between subgroups of patients, patients were


FIGURE 1. Schematic illustration of classification of patients into disease states.


whether they respond to medical treatment or not and can consequently clinically be in remission. Refractory patients are those receiving medical treatment for CD but who are not responding to that treatment. Patients in the state Active have a relatively high disease activity, but are either not receiving treatment or have an unclear response to the treatment they are receiving. The Surgery state includes patients who have recently undergone an operation.

Statistical Analysis classified into mutually exclusive disease states: Remission, Response, Active, Refractory, and Surgery. The classification of patients was done according to a combination of an assessment of current disease activity by the types of current or recent medical or surgical therapy that the patients had received and by the patients’ response to medical therapy. Current treatment and response to treatment has previously been used to stratify CD patients.6,13 Disease activity was assessed using the modified HBI, and remission was defined as an HBI score of less than 4.7, which corresponds to CDAI remission.28 The disease states were defined as follows, which is also illustrated in Figure 1: Response: Patients who were on the medications peroral corticosteroids, immunosuppressive drugs, or anti-TNF therapy with the result remission or response. Patients who responded to some drugs but were refractory to others were assigned to the response state if their HBI was lower than 4.7 and to the refractory state if their HBI was higher than 4.7. Refractory: Patients who were on medications such as corticosteroids, immunosuppressive drugs, or anti-TNF therapy without response or had been on such medication within a certain period of time (2 months for corticosteroids and 6 months for immunosuppressive drugs and anti-TNF therapy) without response. Also, patients who had been on corticosteroids for more than 6 months and still had an HBI over 4.7 were assigned to the refractory state. Surgery: Patients who had undergone an operation and who were discharged from the operation-related hospital admission within 6 weeks prior to the time of the interview. Remission: Patients with an HBI lower than 4.7 who were not assigned to any of the above-mentioned states. Active: Patients with an HBI higher than 4.7 who were not assigned to any of the above-mentioned states and patients who had an unclear response to medical treatment. Hence, in this classification of patients Remission refers to patients with low disease activity and with no need for treatment with corticosteroids, immunosuppressive drugs, or anti-TNF therapy. Response comprises patients who are treated with corticosteroids, immunosuppressive drugs, or anti-TNF therapy and who are responding to that treatment. These patients are classified according to

Since the underlying sampling distribution of both HBI, costs of care, and HRQoL is nonnormally distributed, confidence intervals were estimated using bias-corrected and the accelerated bootstrapping method.29,30 To test for differences in HBI, costs, and health utility between disease states the nonparametric Kruskal–Wallis test was used.31 To identify determinants of costs of care, regression analysis was carried out using demographic characteristics and different disease characteristics as predictors. Since the predictors might impact differently on different types of costs, the model was estimated both on total costs and on direct and indirect costs separately. Due to cost data being skewed to the right, the regression models were estimated using a log-link Generalized Linear Model (GLM) assuming a gamma-shaped distribution of the dependent variable.32 All analyses were performed using the statistical package STATA 9.2 for Windows (Statsoft, Tulsa, OK), and statistical significance was evaluated at the 5% level.

RESULTS Patient Demographics and Disease Indicators A total of 667 patients who fulfilled the inclusion criteria were selected from the patient lists at the study centers. Of these, 152 of the patients (23%) could not be contacted, either due to patients not responding to study invitation or to old contact details in the patient list. Out of the 515 patients who were contacted for study enrolment, 420 (82%) participated in the study, 59 patients (11%) chose not to participate, and 36 (7%) were unable to participate for various reasons. Table 2 shows the characteristics of the 420 participating patients. In Table 3 disease activity, as measured by HBI, is presented for the different disease states. Twenty-seven patients had missing data on the number of loose stools and complete data on all other components of HBI. For these patients the number of loose stools per day was imputed using the other components in the HBI in a single regression imputation model. Four patients had incomplete data on several components of HBI, and consequently could not be classified into 1 of the disease states. Among the remaining 416 patients the majority were in the states Remission and Response (41% and 29%, respectively). With only 2 patients

1885


Mesterton et al

TABLE 2. Demographic and Disease Characteristics of the Study Population (n¼420)

Mean age in years Male (%) Employed (%) Mean disease duration in years Mean Harvey-Bradshaw Index Disease involvement ileum (%) Disease involvement colon (%) Disease involvement other (%) Presence of fistulae (%) Presence of anal fissure (%) Presence of ileostomy (%)

Mean (SD)/ Proportion

Range

50.2 (15.1) 45.6% 64.1% 20.0 (13.2) 4.5 (4) 71.9% 71.7% 1.4% 20.2% 6.2% 11.3%

19-89 0-1 0-1 1-54 0-27 0-1 0-1 0-1 0-1 0-1 0-1

being classified into the Surgery state, data for this group must be interpreted with caution. Patients in Remission had an average HBI of 1.83, compared to 8.88 for patients in Refractory. The differences between patients in Remission, Response, and Active were all statistically significant (P < 0.001 in all cases), while patients in Active and Refractory had similar disease activity.

Resource Use Resource utilization is presented in Table 4, both in different stages of CD (the Surgery group is not included in the table due to the low number of patients in that group) and for the whole sample. Twenty-six patients (11%) had been admitted to inpatient care during the last 12 months. The average number of hospital days per patients among those who had been admitted at least once was 9.1 days. Except for visits to a specialist physician and nurse, outpatient visits were very rare. The most common medical treatments were immunosuppressive drugs and 5-ASA (29% and 28%, respectively). Eighteen patients (4.3%) were treated with immunomodulators (TNFa-inhibitors) at the time of the study visit. A total of 21% of the patients had undergone endoscopy during the last 12 months. With regard to the difference in resource use between disease states, the overall pattern is that the difference was greater in terms of the proportion of patients using each resource, rather than the amount of resources used by those patients. Hospitalizations were most frequent among refractory patients. However, among patients admitted to inpatient care, responding patients had the highest number of hospital days on average. Among the refractory patients 11% had visited an emergency care unit during the last 3 months. Also for the other types of outpatient visits, the frequency of utilization was highest among refractory patients, albeit with

1886

no striking differences. The use of the various medications was least common among patients in remission and most common among patients in the refractory and response states, which is not surprising given their higher disease activity and the fact that those 2 states are partly defined by current medical treatment. The use of diagnostic tests did not differ markedly between disease states.

Costs The estimated 4-week cost in different disease states, divided into different categories of costs, is presented in Table 5. The estimated cost in the 4 main disease states is also illustrated in Figure 2. The average 4-week cost for this sample of CD patients was €721 (USD 988), corresponding to a yearly cost of around €9375 (USD 12,844). Almost two-thirds of the total costs were due to lost productivity. Of direct medical costs of care, 45% were due to hospitalizations and 30% were due to pharmaceuticals. The distribution of costs within the sample was highly skewed; the 20% of patients with the highest costs accounted for 78% of all costs incurred by the sample. The total costs varied by a factor 66 between patients in remission and those in the surgery state and by more than a factor of 5 between refractory patients and patients in remission. Pairwise tests showed that patients in all of the more severe disease states had higher costs than patients in remission (P < 0.001). The patients in remission had lower costs than the other disease states in all cost categories, except for complementary medicine. The difference in total cost between Response and Active was not statistically significant; the notable difference between the 2 was that costs of medical treatment were much higher in Response and that productivity costs were higher in Active. Refractory patients had higher costs than patients in the states Active and Response (P ¼ 0.01 and P ¼ 0.04, respectively). For all disease states except Surgery, productivity costs accounted for a majority of the estimated total cost. For most cost items, the states Active and Refractory were relatively close; the large difference in total cost between them was almost entirely due to higher costs of medication in the latter group. TABLE 3. Disease Activity in Different Disease States State

n

Mean HBI

95% CI

Range

Remission Response Active Refractory Surgery All patients

169 121 97 27 2 416

1.83 4.20 8.20 8.88 10.68 4.51

1.63-2.04 3.65-4.90 7.60-9.01 7.74-10.53 9.57-11.79 4.13-4.91

0-4.57 0-18.71 4.71-27 4.86-18 9.57-11.79 0-27



TABLE 4. Proportion of Patients Using Each Type of Medical Resource and Average Resource Use Among Patients Using Each Resource Remission (n¼169) Type of Resource (Recall Period) Diagnostics (12 months) Endoscopy Abdominal x-ray CT scan MRI Inpatient care Hospitalization, days per patient (12 months) Outpatient visits (3 months) Emergency care unit GP Specialist physician Nurse Occupational therapist Ergonomic therapist Psychologist Drugs at the time of study Peroral 5-ASAs Rectal 5-ASAs Peroral corticosteroids Rectal corticosteroids Immunosuppressive drugs Anti-TNF therapy Antibiotics

Response (n¼121)

Active (n¼97)

Refractory (n¼27)

All (n¼420)

%

Mean*

%

Mean*

%

Mean*

%

Mean*

%

Mean*

16.0 3.6 2.4 4.1

1.1 1.7 1.3 1.0

20.7 2.5 5.8 5.0

1.3 1.0 1.6 1.0

23.7 5.2 5.2 10.3

1.4 1.6 1.2 1.2

48.1 7.4 7.4 18.5

1.2 1.0 1.5 1.2

21.4 4.0 4.3 7.1

1.3 1.4 1.4 1.1

7.7

7.2

7.4

14.0

14.4

7.4

25.9

11.7

11.0

9.1

1.2 0.6 16.6 12.4 0.0 0.0 0.0

1.5 1.0 1.1 1.5 0.0 0.0 0.0

3.3 0.8 25.6 40.5 2.5 0.0 0.0

1.0 1.0 1.4 2.2 2.0 0.0 0.0

6.2 0.0 21.6 32.0 1.0 0.0 2.1

1.5 0.0 1.4 3.6 1.0 0.0 1.0

11.1 0.0 37.0 48.1 3.7 0.0 0.0

1.3 0.0 1.1 2.3 2.0 0.0 0.0

3.6 0.5 21.9 28.1 1.2 0.0 0.5

1.3 1.0 1.3 2.5 1.8 0.0 1.0

27.5 1.8 9.1 0.0 0.6 0.0 2.4

— — — — — — —

30.5 0.8 20.5 0.8 81.8 10.8 4.2

— — — — — — —

22.7 7.2 6.6 1.0 6.3 0.0 2.1

— — — — — — —

29.6 0.0 42.1 0.0 37.0 14.8 11.1

— — — — — — —

27.7 2.6 13.9 0.5 28.7 4.3 3.3

— — — — — — —

*Average use of the resource by the patients using this resource during the recall period.

The results of the regression analysis of the determinants of costs are presented in Table 6. The by far strongest predictor of costs in all 3 models is the HBI; a 1-point increase in the index increases total costs by 25%. The impact of an increase in HBI is stronger on indirect costs

(32%) than on direct medical costs (17%). Both types of costs are decreasing with age; a 1-year increase in age decreases indirect costs by 4.2% and direct costs by 1.9%. It is hard to determine to what extent this is an effect of the age of the patient or of the disease duration, which are

TABLE 5. Estimated 4-week Cost (€) in Different Disease States (2007 Exchange Rate €1¼USD 1.37)23

Inpatient treatment Drugs Outpatient treatment Diagnostics/Imaging Complementary medicine Productivity costs Total 95% CI for total costs

Remission (n¼169)

Response (n¼121)

Active (n¼97)

Refractory (n¼27)

Surgery (n¼2)

All (n¼420)

22 14 17 10 17 176 255 189-363

85 152 46 13 8 526 831 647-1096

41 18 53 19 7 753 891 680-1116

108 315 63 31 10 834 1360 918-2125

13181 649 148 48 0 2959 16,984 16,813-17,156

116 77 37 15 11 465 721 602-905

1887


Mesterton et al

DISCUSSION

FIGURE 2. Estimated 4-week cost (€) in the states Remission, Response, Active, and Refractory. For illustration purposes the Surgery group was excluded from the figure.

obviously highly correlated. The effect of age on indirect costs is most likely due to lower employment rates in older patients. Regarding the effect of age on direct costs, this could be due to the fact that direct medical costs decrease with longer disease duration.13,33 Another statistically significant effect was that the presence of fistulae almost doubles the direct medical costs related to CD.

Health Utilities Health utilities ranged between 0.55 and 1 in the sample, with an average of 0.88. In terms of different stages of CD, Surgery-patients had the lowest average health utility at 0.77, although with only 2 patients it is hard to draw definite conclusions about this group. Among the other states utility differed significantly, being 0.92 in Remission, 0.90 in Response, 0.82 in Active, and 0.81 in Refractory. Pairwise differences between the states Remission, Response, and Active were all statistically significant (P < 0.05 in all cases), while patients in the states Active and Refractory reported very similar utilities.

CD is a chronic disease that appears early in life, causing substantial morbidity over the course of patients’ lives and entailing very high societal costs, both in terms of direct medical costs and in terms of lost productivity. Our observational study clearly demonstrates that by far the strongest predictor of societal costs for CD is the disease activity. Patients in the state Remission, who had a low disease activity and no need for active medical treatment to maintain remission, had 3–5 times lower costs than for patients with a more active disease. Furthermore, even among patients who were not in the state Remission, the patients who were responding to treatment had significantly lower costs of care and higher quality of life than patients who had high disease activity but were not treated or did not respond to treatment. This emphasizes the potential economic benefit of treatments that can induce and maintain remission in patients with high disease activity. Both in terms of the proportion of costs of CD that are due to indirect costs, direct costs, and hospitalizations and in terms of the distribution of costs among patients, the results of this observational study corroborate many previous findings on the economic burden of CD.7–9,34–36 In our study, indirect costs of care were estimated at 64% of total costs, even though labor force nonparticipation and loss of leisure time were not included. However, as opposed to previous studies, this study included the costs of reduced productivity while at work in the estimates of indirect costs. Hospital costs have previously been estimated at 53%–63% of total direct medical costs in Europe, and here the proportion was estimated at 45%. One reason for the lower proportion of hospital costs in our study could be that the relatively recent introduction of the effective and expensive TNF-a inhibitors has increased the proportion of direct costs that are due to pharmaceutical

TABLE 6. GLM Regression of Determinants of Different Types of Costs (n¼414) Total Costs

Age Male Harvey-Bradshaw Index Disease involvement ileum Disease involvement colon Presence of fistulae Presence of anal fissure Presence of ileostomy Constant

Direct Costs

Exp(b)*

P-value

Exp(b)*

P-value

Exp(b)*

P-value

0.969 1.192 1.247 1.242 1.014 1.221 0.959 1.227 636