Abstract
Background
Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma. Cure rate is high after first-line treatment, however, prognosis for relapse/refractory patients remains poor. Multiple treatments to DLBCL have recently been approved, therefore it is important to understand the underlying total costs of the different treatment lines.
Objective
To investigate the total healthcare costs per line of DLBCL treatment (first-line (1LoT), second-line (2LoT), third-line (3LoT), including best supportive care (BSC)), and to estimate patients’ labour market attachment and sickness benefit level.
Method
A nationwide real-world registry study covering the Danish DLBCL patients diagnosed 2012-2021 (August) from the National Lymphoma Registry (LYFO). Data were linked with other national registries to estimate the total costs per treatment line. Additionally, costs per line were adjusted to costs per-patient-per-month (PPPM).
Results
4,159 DLBCL patients were included. Average progression-free time before progression to next LoT or being cured decreased from 2 years (1LoT) to half a year (3LoT + BSC). Average total healthcare costs per patient were estimated to €51,660 (1LoT), €59,094 (2LoT) and €35,808 (3LoT + BSC). Adjusting for progression-free time periods, costs per PPPM increased with later treatment lines. Patients’ labour market attachment was in general low, and weeks of sickness benefit increased with later treatment lines.
Conclusion
The study suggests that the total costs and weeks of sickness benefit increase with later treatment lines suggesting that more effective first-line treatment options not only have an importance for the patient prognosis but also may have an economic advantage for the healthcare sector and society.
Keywords: Diffuse large b-cell lymphoma, Lines of treatment, Real-world data, Total costs, Labour market attachment
Background
Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma (NHL) [1] and has an incidence rate of 8 patients per 100,000 population [2]. A German claims data study estimated the prevalence and cumulative incidence of DLBCL in 2020 to be 36.7 and 12.7 patients per 100,000 population, respectively [3]. The risk of DLBCL increases with age and is generally higher in males than females.
The standard treatment of DLBCL in first-line treatment (1LoT) is rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). In some countries, like in Denmark, high-risk patients 18–80 years of age are treated with different R-CHOP regimen or for patients 18–65 years of age without comorbidities in an intensified R-CHOEP regimen (CHOEP: Chemotherapy with cyclophosphamide, doxorubicin, etoposide, vincristine and prednisone) [4;5]. Elderly patients >80 years can still be eligible to curatively intended treatment with dose attenuated R-CHOP, if they are fit and without significant comorbidity [4, 6]. Patients with relapsed or refractory (R/R) DLBCL are offered a number of treatment alternatives in second line (2LoT) and in third line (3LoT), including best supportive care (BSC) [4]. Treatment options depend on age and co-morbidities. In 2LoT, depending on the time of R/R, fit patients eligible for autologous stem cell transplantation (ASCT) can be treated with curatively intended treatments with either salvage chemotherapy followed by ASCT [7] or chimeric antigen receptor (CAR) T-cell therapy, which is approved for a subset of this patient group [4]. There is no standard 2LoT for elderly (>80 years), frail and ASCT ineligible patients, why they are treated with BSC, inclusive steroids, different chemotherapy regimens or palliative care only [4].
In an extensive epidemiological review from 2022 the 5-year overall survival (OS) rate for all DLBCL patients ranged from 53.1% to 61.6% [8]. However, among R/R patients the 2-year OS rate was as low as 13.9% to 16.0%. This low OS rate was also reflected in a Danish population-based study which showed that DLBCL patients who had received three or more lines of therapy had a median OS of 5.8 months [9]. Because of the poor prognosis for patients with R/R DLBCL, much effort is invested in the development of new therapies for DLBCL, both in the R/R setting, but also in front line to improve outcomes and reduce number of patients’ progression to subsequent lines.
Understanding the outcomes of current available therapies in the different lines of therapy of DLBCL as well as the total costs of DLBCL treatment is essential for the healthcare sector, the payers and the patients. It will show the economic implications of being in a specific line of DLBCL treatment and guide correct prioritization of resources and optimal treatment of the individual patient. Furthermore, severe diseases like DLBCL are expected to have wider implications beyond the healthcare sector such as the patients’ ability to actively go to work versus being on sick leave and receiving sickness benefit, both of which are important topics to investigate.
A number of studies have investigated the healthcare utilization costs or cost of illness for DLBCL from a healthcare sector perspective. However, the majority of these studies were based on data the US, Japan, Canada and Korea with only a few from Europe (Germany, Switzerland and Portugal) [10–13]. Additionally, none of these investigated the impact beyond the healthcare sector in terms of labour market attachment and use of sickness benefit. Furthermore, the Real World (RW) data sources for many of these studies were based on non-nationwide, but selected populations relying on data from either claims databases, single centers or regions or specific cohorts (e.g. US Veterans) [11, 13–16]. Only one study using Korean Health Insurance data utilized data covering the entire population of the nation [17].
Denmark has a tax-financed universal healthcare system and more than 200 different registries and clinical databases that covers all Danes and of which some have more than half a century of longitudinal data [18]. These national registries and databases are of high quality and can be combined on an individual level by the existence of a unique identifying number from the Civil Registration System (CPR number) that every Danish citizen has. The national registries and clinical databases covering RW data can be used for research purposes. Combining data from the different registries has been used to track illnesses in the Danish population as well as assessing effect, safety, and medicine utilization [19]. One of the clinical databases is the Danish National Lymphoma Registry (LYFO) established in 1982, and being nationwide from 2000, aims at monitoring and improving the diagnostic evaluation and the quality of treatment of all lymphoma patients in Denmark, including DLBCL patients [20]. This LYFO registry can be combined with other relevant Danish national registries to investigate the DLBCL disease further.
The objectives of the present study were, from a healthcare sector perspective, to estimate and compare the total costs of treatment in the different treatment lines of DLBCL, based on RW data covering all DLBCL patients in Denmark, as well as to assess the broader societal impact of the disease in terms of the level of patients’ labour market attachment and use of sickness benefit.
Methods
The study was designed as a nationwide registry-based retrospective cohort study, where the different outcomes were compared for 1LoT DLBCL patients with DLBCL patients in the later treatment lines of 2LoT and 3LoT + BSC. Patient registry data from the Danish LYFO registry [20] was linked with a number of other relevant national registries in Denmark containing data on health-resource utilization and labour activity. Data on costs related to the cancer medicine prescribed and administered to DLBCL patients at the hospital was, however, not included in the analyses as explained further below.
Data sources
An overview of all registries used in the current study is presented in Table 1.
Table 1.
Overview of Danish registries used, types of data, and data time period
| Registry | Type of data | Data time period |
|---|---|---|
| National Lymphoma Registry (LYFO) | Data on the Danish DLCBL patients | 2012–2022 |
| Civil Registration System (CRS) | Linkage data with the CPR number – unique number that all Danes have | 2012–2022 |
| National Patient Registry (NPR) | Activity at the hospital level (inpatient and outpatient activity, diagnose codes (ICD-10)) | 2012–2022 |
| Diagnosis Related Groups (DRG) | The DRG rates express the hospitals’ average operating costs for activities. | 2012–2020 |
| National Health Service Registry (SSSY – Sygesikringen) | Health services provided in the primary healthcare, e.g. visits to general practitioner. | 2012–2021 |
| Registry of Pharmaceutical Sales (LMDB) | Pharmaceutical sale in the primary care setting (primary care medicine), including the public reimbursement part of primary care medicine. | 2012–2021 |
| Labour Market Attachment without standardization of hours (AMR-UN) | The Danish populations labour market status and attachment (without hourly rationing) | 2012–2021 |
| Labour Market Attachment with standardization of hours (AMR-TN) | The Danish populations labour market status and attachment (with hourly rationing) | 2012–2021 |
| Social pensions (SOCP) | Data on different types of social pensions given due to illness, early retirement, public pension, etc. | 2012–2021 |
| DREAM database | Information on labour market affiliation, public payment transfers and disability payment (sickness benefits), etc. | 2012–2022 |
The primary data source to identify the Danish DLBCL patients was the LYFO registry [20]. LYFO is a nationwide registry which to date covers approximately 95% of all Danish lymphoma patients [21]. In LYFO all Danish hematology departments are obliged to register the following information: all new patients at the time of diagnosis; primary treatment at the end of 1LoT; reason and time of R/R, choice of 2LoT as well as status at follow-up or death. As the LYFO registry does not contain data on patients in 3LoT, the outcome of 3LoT in the present study is therefore not based on a clinical evaluation, but on the assumption that the clinical response evaluation after 2LoT categorized as progressive disease or no change to progression from 2LoT and thus to 3LoT, including BSC. Patients initially achieving complete remission in response to 2LoT, but which later developed into R/R, were as a result of the described approximation not included in this group.
The LYFO registry was cross-referenced with the Danish National Patient Registry and the Danish National Pathology Registry. Additionally, a linkage to the Danish Civil Registration System was made to obtain data on date of death [20]. In the end this secures that no patient is lost to follow-up.
The Civil Registration System Registry (CRS) contains data for every Dane on date of birth, gender, residence, family relations, etc [22]. The National Patient Registry (NPR) includes data on all patient contacts at the hospital level (inpatient and outpatient care in the hospital), including the diagnosis (ICD-10 code system), type of care and its timing and duration [23]. The costs of the hospital activity were estimated based on the Diagnosis Related Groups (DRG) tariff system. This system of DRG express the hospitals’ average operating costs for activities within a specific disease and are considered good proxies of the hospital costs. The National Health Services Registry (SSSY) contains data on the health services provided to patients in the primary healthcare setting [24]. The Registry on Pharmaceutical Sales (LMDB) provides data on all sales of pharmaceuticals in the primary care setting prescribed by general practitioners or private practicing specialist and sold by retail pharmacies, including the public reimbursement part of primary care prescription medicine [25]. Hospital medicine, including cancer medicine, administered at the hospital and delivered from the hospital pharmacy, are not part of this registry. The registries of labour market attachment (AMR-UN and AMR-TN) are registries of labour market attachment, covering citizens with full- or part-time occupation. The Registry on Social Pensions (SOCP) is a registry of all types of social pensions given to Danish residents due to retirement, illness, or other circumstances. The DREAM database is an additional labour market database, which includes information on weekly labour market public transfer payments, i.e., unemployment benefits or disability payments, membership of unions, etc. for all Danish citizens since 1991 [26].
Data on hospital medicine, including hospital-administered cancer medicine, can be found in the Danish Registry on Hospital Medicine Use. However, this registry is relatively new and only includes cost data based on official list prices. Additionally, it only covers a subset of the Danish population from 2018 onward. As a result, the data on utilization and costs of cancer medicine was deemed insufficient for the data period of 2012–2022 of the present study.
The extracted data were obtained from patients diagnosed in the period 2012–2022, which were also the data time period used for data coming from the other linked registries, although in some cases the last year or two were missing (see Table 1).
Patient population
Individuals diagnosed with ICD-10 code C83.30 DLCBL) in the period 2012–2022 were retrieved from the LYFO registry. Patients were included if they fulfilled the following two criteria: (1) Adults ≥ 18 years diagnosed with DLBCL between 2012 and 2021, i.e. being newly diagnosed with DLBCL no later than end of August 2021, who had received at least one line of treatment for DLBCL and (2) A DLBCL histology, including DLBCL not otherwise specified (NOS), High-grade B-cell lymphoma (HGBL), and Primary mediastinal B-cell lymphoma (PMBCL). Patients were excluded, if they fulfilled at least one of the following criteria: (1) Primary central nervous system (CNS) lymphoma or CNS involvement at the time of diagnosis; (2) Richter transformation; (3) Post-transplant lymphoproliferative disorder or (4) Transformed low grade lymphoma.
Patient characteristics and treatment responses were in the study described for the 1LoT, 2LoT and 3LoT + BSC treatment groups. Clinical response, i.e. relapse and refractory, to the treatment was assessed according to the Lugano classification criteria [27], which are applied in clinical practice in Denmark. Accordingly, these criteria were used to define clinical response in the data obtained from the Danish LYFO registry [20]. For the analyses, the patient groups were defined as follows: The 1LoT patient group, which is the overall DLBCL patient group containing all patients ≥ 18 years who had been diagnosed with DLBCL and received at least one line of treatment; the 2LoT patient group which consisted of patients who suffered a relapse or were refractory to the first line treatment (R/R) and had a second line of treatment initiated; and the 3LoT + BSC patient group, which is a group of patients whose treatment response was evaluated as either progressive or with no change after 2LoT following the definition presented above.
Hence, the three patient groups were nested within each other, such that the 3LoT patients prior to their progression were initially part of the 2LoT patient group and the 2LoT patients were initially part of the 1LoT patient group, also prior to their progression. Since the study did include patients diagnosed with DLBCL up until the end of August 2021 who received at least one line of treatment, some patients have had a maximum of 1.5 years of follow-up data and may not yet progressed to later lines of therapy. As a result, the analysis of average costs for each treatment line was based on the patients available within each respective line.
Data analysis
The analysis of the total healthcare costs based on the registry data, focused on healthcare resources and costs associated with hospital inpatient care and outpatient contacts, and primary care medicine consumed including patients’ out-of-pocket costs. Cancer medicine administered at the hospital level was not included and is not included in the estimation of the total healthcare costs with the reasons provided above. All contacts and utilization were counted per patient during the progression-free time in each line of treatment and the costs per patient for the specific resource utilization, were then summed up to be able to calculate the average costs per patient. Labour market attachment was counted as the average number employment days during the time period being in each LoT. Similarly, sickness benefit was calculated as the number of weeks with sickness benefit.
All results of the three lines of DLBCL treatment were estimated and reported as observed resource utilization and costs, i.e. their actual total utilization and associated costs. Costs in each of the three lines of treatment were calculated as the average costs based on the available data for patients in the specific line. Furthermore, to be able to compare the different lines of DLBCL treatment and to take differences in treatment periods into account, the resource utilization was standardized to yearly estimates, i.e. 365 days, for all three lines. The costs of the three different treatment lines were on the other hand adjusted (standardized) to the costs per-patient-per-month (PPPM) following the same method as presented in other studies in the field, e.g. Borchmann et al. [10] and Gatwood et al. [14]. The statistical analyses were performed using R Studio.
Table 2 shows the definition of the different cost and activity variables.
Table 2.
Description of activity and costs variables
| Overall variable | Specific variable | Definition |
|---|---|---|
| Hospital costs | Bed days – inpatient | Number of days of hospitalization |
| Contacts – outpatient | Number of contacts to hospital ambulatory (outpatient) | |
| Costs – inpatient | Total cost of hospitalization including both bed day costs and treatment costs. | |
| Costs – outpatient | Total cost of outpatient visits | |
| Medicine costs, primary care medicine | Patient payment | Out of pocket payment |
| Subsidized price (by region/health insurance) | Reimbursement by public insurance | |
| Total payment (price incl. VAT etc.) | Total cost of prescription medicine | |
| Employment | Days of employment (standardized Full Time Equivalent (FTE)) | Employment (only individuals with > 0 days of employment during treatment) |
| Sick benefits | Weeks of sick benefits | Number of weeks (min. 1 day a week) during treatment |
Ethical and regulatory approval
Following Danish law, approval from the Danish Data Protection Agency and the Ethics Committee in Denmark was not required for registry-based studies. Furthermore, being a non-interventional study the study was not covered by the law on pharmaceutical interventional studies and did neither require approval by the Danish Medicines Agency.
Results
Patient demographics and treatment response
A total of 4,159 patients over the study period of 10 years (2012–2022) were included in this retrospective study. The median age was 69–70 years in the three treatment groups with an overall age range of 18–101 years. More males (57%) than females (43%) were diagnosed with DLBCL, and a higher fraction of males compared to females progressed to 2LoT (65%) and 3LoT (68%) (Table 3).
Table 3.
Patient demographics
| 1LoT (n = 4,159) | 2LoT (n = 518) | 3LoT (n = 161) | |
|---|---|---|---|
| Age, years (median; range) | 70 (18–101) | 69 (18–93) | 70 (19–93) |
| Females (number; %) | 1,782 (43%) | 182 (35%) | 52 (32%) |
| Males (number; %) | 2,377 (57%) | 336 (65%) | 109 (68%) |
Of the 4,159 patients initiating treatment in 1LoT, 85.8% (3,570 patients) were treated with a curative intent (R-CHOP, R-CHOEP or R-Mini-CHOP) (Table 4). For R/R patients progressing to 2LoT, only 20.3% (105 patients of the 2LoT) were treated with a curative intent using high dose chemotherapy with ASCT.
Table 4.
Treatment response in the three lines of treatment
| Clinical assessment | 1LoT (n = 4,159)1 | 2LoT (n = 518)1,3 | 3LoT (n = 161) |
|---|---|---|---|
| Patients with relapse or being refractory (R/R) (number; %) | 601 (14.0%)2 | 161 (31.0%) | Not available in LYFO |
| Patients achieving Complete Response (number; %) | 3,069 (73.8%) | 197 (38.0%) | Not available in LYFO |
| Patients achieving partial Response (number; %) | 225 (5.4%) | 48 (9.3%) | Not available in LYFO |
| Patients with progressive disease (number; %) | 212 (5.1%) | 135 (26.1%) | Not available in LYFO |
| Patients with no change (number; %) | 29 (0.7%) | 17 (3.3%) | Not available in LYFO |
| Patients missing confirmed clinical assessment (number; %) | 303 (7.3%) | 44 (8.5%) | N/A |
| Patients with disease-related death (number; %) | 321 (7.7%) | 77 (18.0%) | N/A |
1. The total number of patients in each LoT is the sum of the patients achieving complete response, patients achieving partial response, patients with progressive disease, patients with no change, patients with missing confirmed clinical assessment and patients with disease-related deaths
2. The 601 patients with relapse or being refractory consists of 212 patients with progressive disease, 321 patients with disease-related death and 68 patients with relapse
3. The 518 patients represent those of the 601 patients with relapse or refractory in 1LoT, who fulfilled the retrieval criteria of the study as explained in the Patient Population section
Overall risk of R/R was 14% in 1LoT and 31% in 2LoT (Table 4). Either a complete response (CR) or a partial response (PR) was obtained in 79.4% of 1LoT patients compared to 47.3% of the 2LoT patients. In line with this, progression of disease was naturally more frequent in the 2LoT patients group compared to the 1LoT patient group as shown in Table 4. Confirmed clinical assessment was missed in 7.3% of the patients in the 1LoT group and 8.5% of the patients in 2LoT missed a confirmed clinical assessment.
Total healthcare costs of treatment in different lot
Danish DLBCL patients in 1LoT were on average progression-free for approximately 2 years before progression to 2LoT or otherwise being cured (Table 5). In 2LoT, DLBCL patients were progression-free for 1.3 years on average, and only half a year on average for patients in the 3LoT stage (Table 5).
Table 5.
Observed and standardized resource consumption and labour activities for DLBCL patients in the three lines of treatment
| Observed resource consumption/activity | 1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
|---|---|---|---|---|
| Progression-free time in LoT | Average time as progression-fee before progressing to next LoT due to R/R or being cured (Mean; SD) | 2.04 years | 1.31 years | 0.54 years |
| Hospital | Inpatient days at hospital (Mean; SD) | 24 days (28.7) | 35 days (35.0) | 24 days (24.5) |
| Outpatient contacts (Mean; SD) | 23 days (21.2) | 19 days (22.1) | 15 days (19.3) | |
| Employment |
Days of employment* (Mean; SD) *average of persons with > 0 employment |
464 days (391.1) (n = 1,229 (30%)) |
316 days (332.8) (n = 88 (17%)) |
172 days (212.0) (n = 24 (15%)) |
| Sick benefits |
Weeks of sick benefits* (Mean; SD) *average of persons receiving sickness benefits |
45 weeks (31.7) (n = 836 (20%)) |
42 weeks (42.4) (n = 90 (17%)) |
25 weeks (28.6) (n = 22 (14%)) |
|
Adjusted (standardized) resource consumption/activity - per patient per year |
1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
| Hospital | Inpatient days at hospital (Mean; SD) | 11.8 days (14.1) | 26.8 days (26.7) | 44.7 days (45.4) |
| Outpatient contacts (Mean; SD) | 11.3 days (10.4) | 14.2 days (16.8) | 27.6 days (35.7) | |
| Employment |
Days of employment* (Mean; SD) *average of persons with > 0 employment |
189 days (159.3) (n = 1,229 (30%)) |
182 days (191.5) (n = 88 (17%)) |
151 days (186.2) (n = 24 (15%)) |
| Sick benefits |
Weeks of sick benefits* (Mean; SD) *average of persons receiving sickness benefits |
19 weeks (13.6) (n = 836 (20%)) |
30 weeks (12.0) (n = 90 (17%)) |
33 weeks (37.1) (n = 22 (14%)) |
Table note: SD = Standard deviation
In terms of healthcare resource utilization, a patient in 1LoT did spend on average 24 days being hospitalized during the 2 years of progression-free time, i.e. 3% of the days in the progression-free time period and had 23 outpatient visits to the hospital (Table 5). In comparison, the average DLBCL patient in 2LoT did spend 35 days being hospitalized, i.e. 7% of the days in the progression-free time period in 2LoT, as well as did visit the outpatient clinic in the hospital 19 times. Finally, the average 3LoT DLBCL patient did spend 24 days on average in the hospital. This accounted for 12% of the progression-free time in 3LoT. Furthermore, the average 3LoT patient has 15 visits to the outpatient clinic. This highlights the severity of DLBCL for patients who were R/R to 1LoT, especially when considering that the average time spent in 3LoT was only a quarter of the average progression-free time in 1LoT.
Standardizing the hospital resource utilization assuming one full year (365 days) in each line of treatment also confirmed the increased disease burden in the R/R setting since the 3LoT group had on average four times as many inpatient days in the hospital compared with patients in the 1LoT group (44.7 versus 11.8, Table 5). Similarly, DLBCL patients in the 2LoT group had more than twice as many inpatient days as the patients in the 1LoT group (26.8 versus 11.8) confirmed the increased severity of the disease with each additional line of therapy. A similar pattern was seen for outpatient visits (Table 5).
Table 6 shows the total healthcare costs of treatment in the three different lines of treatment for DLBCL. In accordance with the resource utilization as presented in Table 6 the observed total costs of treatment were highest in the 2LoT group with €59,094 for the average DLBCL patient in that line followed by the total costs in 1LoT of €51,660 and €35,808 in 3LoT. Again, these differences were due to the different progression-free time in each LoT before progression/death with the 3LoT only lasting half a year on average. Hospital costs were the largest contributor to the total costs of treatment with inpatient costs accounting for around two-thirds of the costs. The observed inpatient hospital cost was estimated to €31,358 for 1LoT, €45,968 for 2LoT, and €24,845 for 3LoT, whereas outpatient costs amounted to €19,390, €12,454 and €10,470, respectively, for the three lines of treatment (Table 5). Cost related to primary care medicine was of minor importance.
Table 6.
Observed and standardized healthcare resource utilization for DLBCL patients in the three lines of treatment
| Observed resource utilization/activity | 1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
|---|---|---|---|---|
| Total costs of treatment | Total healthcare costs of treatment (Mean; SD) | 51,660 (51,465) | 59,094 (67,354) | 35,808 (46,882) |
| Hospital costs | Inpatient days at hospital (Mean; SD) | 31,358 (31,388) | 45,968 (48,301) | 24,845 (26,556) |
| Outpatient contacts (Mean; SD) | 19,390 (18,737) | 12,454 (18,039) | 10,470 (19,630) | |
| Medicine costs, primary care medicine | Part reimbursed by the public health services insurance (Mean; SD) | 912 (1,340) | 672 (1,014) | 493 (697) |
| Patient co-payment part (Mean; SD) | 1,163 (1,962) | 876 (1,346) | 663 (1,030) | |
|
Adjusted (standardized) resource utilization/activity - per patient per year |
1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
| Total costs of treatment | Total healthcare costs of treatment (Mean; SD) | 2,055 (2,102) | 3,708 (4,285) | 5,468 (7,235) |
| Hospital costs | Inpatient costs (Mean; SD) | 1,270 (1,282) | 2,921 (3,073) | 3,847 (4,098) |
| Outpatient costs (Mean; SD) | 785 (765) | 791 (1,148) | 1,621 (3,029) | |
| Medicine costs, primary care medicine | Part reimbursed by the public health services insurance (Mean; SD) | 42 (55) | 58 (65) | 80 (108) |
| Patient co-payment part (Mean; SD) | 33 (80) | 44 (86) | 59 (159) | |
Table note: SD = Standard deviation
Due to different average progression-free time periods in each line before cure or eventual progression as presented in Table 6 the total costs and cost components were further adjusted to costs per-patient-per-month (PPPM) for further comparison between the lines. As Table 6 shows the PPPM total costs of treatment in the 3LoT group for the average patient nearly becomes three times as high compared to 1LoT patients and 1½ times as high compared to 2LoT patients (1LoT: €2,055; 2LoT: €3,708; 3LoT: €5,468). Similar patterns were shown for the PPPM hospital costs (inpatient and outpatient) with the exception that the costs of outpatient visits became nearly the same in 1LoT and 2LoT (€785 versus €791). Similar to the resource utilization, the cost estimates demonstrated that the 3LoT group had many inpatient days and contacts with the hospital in a short period of time (half a year).
Labour impact of treatment in different lot
The comparison of the broader societal impact of the DLBCL disease in the different lines of treatment in terms of the level of patient labour market attachment and use of sickness benefit is shown in Table 7. Labour market attachment decreased with increasing LoT (from 30% to 15%, Table 7), but was in general low in all three groups, which is likely due to the average age at disease onset. However, when focusing on individuals who were still part of the labour market (> 0 days of employment) at disease initiation, Table 7 shows that the 1LoT patients had on average 464 days of employment, which means that people in occupation were faced with limited work impairment being in the first line of DLBCL treatment. For the R/R DLBCL patients still working in 2LoT and 3LoT groups the average days of working were 316 days and 172 days, respectively. When employment days were standardized to full treatment years, the days of working for patients were, however, quite similar across all treatment lines (189.1, 181.9 and 151.3 days for 1LoT, 2LoT and 3LoT, respectively). Looking at weeks of receiving sick benefits, patients in 1LoT received 45 weeks compared to 42 weeks in 2LoT and 25 weeks in 3LoT, respectively. Standardizing weeks of sickness benefit, though, showed that the more severe disease the more weeks of sickness benefit needed, i.e. 33 weeks in the 3LoT group and 19.3 weeks in the 1LoT group (Table 7). The share of persons though having sickness benefit though decreased with increasing LoT (from 20% in 1LoT to 14% in 3LoT, Table 7), which was simply a consequence of a decreasing number of people still attached to the labour market in the later LoTs, and therefore not were in scope of sickness benefit.
Table 7.
Observed and standardized labour impact for DLBCL patients in the three lines of treatment
| Observed resource utilization/activity | 1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
|---|---|---|---|---|
| Employment |
Days of employment* (Mean; SD) *average of persons with > 0 employment |
464 days (391.1) (n = 1,229 (30%)) |
316 days (332.8) (n = 88 (17%)) |
172 days (212.0) (n = 24 (15%)) |
| Sick benefits |
Weeks of sick benefits* (Mean; SD) *average of persons receiving sickness benefits |
45 weeks (31.7) (n = 836 (20%)) |
42 weeks (42.4) (n = 90 (17%)) |
25 weeks (28.6) (n = 22 (14%)) |
|
Adjusted (standardized) resource utilization/activity - per patient per Year |
1LoT (N = 4,159) |
2LoT (N = 518) |
3LoT (N = 161) |
|
| Employment |
Days of employment* (Mean; SD) *average of persons with > 0 employment |
189 days (159.3) (n = 1,229 (30%)) |
182 days (191.5) (n = 88 (17%)) |
151 days (186.2) (n = 24 (15%)) |
| Sick benefits |
Weeks of sick benefits* (Mean; SD) *average of persons receiving sickness benefits |
19 weeks (13.6) (n = 836 (20%)) |
30 weeks (12.0) (n = 90 (17%)) |
33 weeks (37.1) (n = 22 (14%)) |
Table note: SD = Standard deviation
Discussion
For more than 20 years, the standard treatment in 1LoT DLBCL had been R-CHOP [28]. However, there remains a need for improved therapies with previous studies showing up to 30–40% of patients with DLBCL with relapse or refractory to front line therapy leading to generally poor survival outcomes [5, 29]. Several recent approvals of novel therapies, highlighted by the recent FDA and EMA approvals of CAR-T cell therapy in the R/R setting, have provided patients with new treatment options, however these treatments are expensive. With many new treatment options, both in development and approved, it is important to understand the healthcare resource utilization as well as the healthcare costs within the different lines of DLBCL treatment.
In the present study, it was found that the resource utilization and costs in the healthcare sector related to treatment of DLBCL increased with each additional line of treatment, when adjusting for the difference in progression-free time in each treatment line. Comparing the observed values without adjusting for time in progression-free periods in each line of treatment the R/R patients in the 2LoT group was on average more costly for the healthcare sector than the 1LoT group, both in terms of inpatient and outpatient costs (Table 6). Standardizing the costs to a cost per-patient-per-month (PPPM), when comparing the three lines of treatment to account for different progression-free time periods before disease progression or eventually cure in each line of treatment, the total costs were nearly twice as high in the R/R 2LoT patient group compared with the 1LoT patient group, whereas the total costs of the R/R 3LoT + BSC patient group were nearly three times as high as the total costs of the 1LoT patient group. These results suggest an increased healthcare burden with additional lines of therapy, albeit being conservative, as the costs of the hospital cancer medicine were not included in the analysis. The approval of novel therapies such as CAR-T are expected to add additional substantial costs to the later treatment lines in the R/R setting and thereby increase the healthcare burden in these lines.
Overall, the results of the study suggested that it is important to consider the total costs of treatment in the different lines of treatment of DLBCL, when considering the best use of scarce healthcare resources. Once DLBCL patients become R/R, i.e. 2LoT or 3LoT, they become more costly for the healthcare sector. Furthermore, the broader societal impact in terms of labour market attachment and receipt and costs of sickness benefit should also be considered.
As described in the background section this study was not the first study to report the total costs of treatment lines in DLBCL. This study was, however, one of only two studies including a full cohort of patients covering a whole nation with no selection bias associated, as well as being the first European study of its kind. The observed costs documented in the present study with a total cost of 1LoT for patients with DLBCL of €51,660, 2LoT of €59,094 and 3LoT of €35,808 – without standardization of different progression-free time periods in the treatment lines – were in line with existing literature, although one should be careful comparing cost results from different countries based on different healthcare systems and practices. For example, the German study by Borchmann et al. [10] based on claims data reported an unadjusted total healthcare cost of €59,868 in 1LoT, €35,870 in 2LoT and €28,832 in 3LoT. Another German single site study by Moertl et al. [11] estimated the total costs of R/R DLBCL to be €44,295 in 2LoT and €32,752 in 3LoT. On the other hand, studies from the US, e.g. Gatwood et al. [14], showed rather higher total cost estimates for 2LoT and 3LoT treatment of DLBCL, which was expected due to differences in healthcare systems and in general higher cost levels between countries. However, importantly all studies demonstrated the same pattern when adjusting for differences in the progression-free time period, because costs increase with increasing lines of treatment, i.e. the average total costs per month for the R/R patients in 2LoT and 3LoT were higher than the costs in 1LoT.
Strengths and limitations
The present study was based on nationwide and linkable Danish national registry data that included all Danish patients in the three lines of DLBCL treatment. Danish national registries were in general considered to be of high quality due to the digitalization of the Danish society as well as the standardized and frequent use of the national registries for research. The data presented in this study was therefore considered as high quality as well as without selection bias in terms of included patients, which sometimes can be the issue with claims data. However, even with the high quality of the data between 7 and 13% missing data were found. An additional limitation of the study was that data for the 3LoT + BSC group were deducted based on the response evaluation in 2LoT as 3LoT patient data are not reported in the LYFO registry. This may have biased the results of the 3LT + BSC group.
The overall risk of R/R may furthermore be underestimated in the 1LoT group, because the study includes patients newly diagnosed in 2021. In cases, where these patients experience a relapse after 2022, this was not included in our data. The actual overall risk of R/R in 1LoT may therefore be higher than the one reported in the study.
Furthermore, the total costs reported in the study, although being from a healthcare sector perspective, did not include the use and costs of cancer medicine administered at the hospital. This was due to the following reasons; (1) the Danish Registry on Hospital Medicine Use only contained data from 2018 and onward, which was found insufficient for the current study covering the period (2012–2022), Additionally, the registry does not yet provide complete coverage of the entire Danish population and (2) the registry only contains official national list prices on hospital medicine and not the actual used net prices, as these are confidential. The actual total costs of the lines of treatment of DLBCL in the present study were therefore most likely underestimated. Overall, the implication of this was expected to mainly affect the later lines of treatment, why the results found in this study can be considered conservative. Future analyses of the total costs of lines of DLBCL treatment should, however, include the cost of the new cancer medicines administered at the hospital level, as these may account for significant part of the total healthcare costs, e.g. between 8 and 19% of the total costs in the study by Borchmann et al. [10].
Finally, no matched historic control group without DLBCL was used for comparison in the analysis of the costs of the DLBCL treatment lines. A historic control group could have identified the costs of all other conditions and diseases not being DLBCL that a patient with DLBCL could also have had as comorbidities. The difference in total costs between the total costs of a DLBCL patient and the total costs of a patient without DLBCL (control) could therefore be regarded as the true costs related to DLBCL. From the analyses in the present study, without a control, it cannot be excluded that some of the healthcare activities (inpatient and outpatient contacts) are related to comorbidities leading to a potential overestimation of the total costs in each treatment line. With the primary aim of comparing the three lines of treatment with each other in the present study, the challenge of leaving out a control group may, however, be of less importance as assuming that all three treatment lines incorporate similar level of comorbidity.
Conclusion
This study showed that total healthcare costs associated with each line of treatment for DLBCL, based on nationwide Danish registry data, increased with each subsequent line. Specifically, total costs in 3LoT + BSC were significantly higher than in 1LoT. Similarly, the proportion of patients receiving sickness benefits rose with later treatment lines, while overall labour market attachment remained low. These findings suggest that more effective first-line treatment options were not only important for patient prognosis but may also offer economic benefits for both the healthcare system and society.
Acknowledgements
Judit Mészáros Jørgensen, Aarhus University Hospital, Peter de Nully Brown, Copenhagen University Hospital and Thomas Stauffer Larsen, Odense University Hospital are acknowledged for their advice and consultancy work during the study. Peter Bo Poulsen, Roche Pharmaceuticals a/s is acknowledged for medical writing work on the final manuscript.
Funding
The study was funded by Roche Pharmaceuticals a/s.
Declarations
Competing interests (disclosure)
Christian Graves Beck and Maya Friis Kjærgaard are employees of Roche Pharmaceuticals a/s. Tove Holm-Larsen and Martin Lucas Jørgensen were paid vendors to Roche Pharmaceuticals a/s. Søren Ramme Bro was a scientific consultant during the study, but did not receive any remuneration for this task.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Li, Y., Wang, Y., Wang, Z., Yi, D., Ma, S.: Racial differences in three major NHL subtypes: Descriptive epidemiology. Cancer Epidemiol 39(1), 8–13 (2015). 10.1016/j.canep.2014.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Swerdlow, S., Campo, E., Pileri, S., Harris, N., Stein, H., Siebert, R., et al.: The 2016 revision of the world health organization classification of lymphoid neoplasms. Blood 127(20), 2375–2390 (2016). 10.1182/blood-2016-01-643569 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pacis, S., Bolzani, A., Heuck, A., Gossens, K., Kruse, M., Fritz, B., et al.: Epidemiology and Real-World treatment of incident diffuse large B-cell lymphoma (DLBCL): A German claims data analysis. Oncol. Ther 12(2), 293–309 (2024). 10.1007/s40487-024-00265-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Diffust storcellet B-celle lymfom. Version 3.0. 5: Sekretariatet for Kliniske Retningslinjer på Kræftområdet (2023). https://www.dmcg.dk/siteassets/kliniske-retningslinjer---skabeloner-og-vejledninger/kliniske-retningslinjer-opdelt-pa-dmcg/dlg/dlg_storcelle-diffet-b-celle-lymfom_v.3.0_admgodk051223.pdf)
- 5.Gang, A., Strøm, C., Pedersen, M., d’Amore, F., Pedersen, L., Bukhet, A.: al. R-CHOEP-14 improves overall survival in young high-risk patients with diffuse large B-cell lymphoma compared with R-CHOP-14. A population-based investigation from the Danish Lymphoma Group. Ann Oncol. 23(1), 147–153 Https:// (2012). 10.1093/annonc/mdr058 [DOI] [PubMed]
- 6.Al-Sarayfi, D., Brink, M., Chamuleau, M., Brouwer, R., van Rijn, R., Issa, D., et al.: R-miniCHOP versus R-CHOP in elderly patients with diffuse large B-cell lymphoma: A propensity matched population-based study. Am. J. Hematol. 99(2), 216–222 (2024). 10.1002/ajh.27151 [DOI] [PubMed] [Google Scholar]
- 7.Gisselbrecht, C., Glass, B., Mounier, N., et al.: Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. Journal of Clinical Oncology 28(27), 4184–4190 (2010). 10.1200/JCO.2010.28.1618 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Garg, M., Takyar, J., Dhawan, A., Saggu, G., Agrawal, N., Hall, A., et al.: Diffuse large B-Cell lymphoma (DLBCL): A structured literature review of the Epidemiology, treatment Guidelines, and Real-World treatment patterns. Blood. 140(Supplement 1), 12106–12107 (2022). 10.1182/blood-2022-169045 [Google Scholar]
- 9.Al-Mashhadi, A., Jakobsen, L., Brown, P., Gang, A., Thorsteinsson, A., Rasoul, K., et al.: Real-world outcomes following third or subsequent lines of therapy: A Danish population-based study on 189 patients with relapsed/refractory large B-cell lymphomas. Br. J. Haematol. 204(3), 839–848 (2023). 10.1111/bjh.19201 [DOI] [PubMed] [Google Scholar]
- 10.Borchmann, P., Heger, J., Mahlich, J., Papadimitrious, M., Riou, S., Werner, B.: Healthcare resource utilization and associated costs of German patients with diffuse large B-Cell lymphoma: A retrospective health claims data analysis. Oncol. Ther. 11(1), 65–81 (2023). 10.1007/s40487-022-00211-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Moertl, B., Dreyling, M., Schmidt, C., Hoster, E., Schoel, W., Bergwelt-Baildon, M., et al.: Inpatient treatment of relapsed/refractory diffuse large B-cell lymphoma (r/r DLBCL): A health economic perspective. Clinical Lymphoma Myeloma and Leukemia 22(7), 474–482 (2022). 10.1016/j.clml.2021.12.018 [DOI] [PubMed] [Google Scholar]
- 12.Trottmann, M., Blozik, E., Hilbig, M., LoVerdi, D., Pedruzzi, M., Scherer, T., et al.: Real-world expenditures and survival time after CAR-T treatment for large B-cell lymphoma in switzerland: A retrospective study using insurance claims data. Swiss Med. Wkly. 153, 3441 (2023). 10.57187/s.3441 [DOI] [PubMed] [Google Scholar]
- 13.Chacim, S., Monjardino, T., Cunha, J., Medeiros, P., Redondo, P., Bento, M., et al.: Costs, effectiveness, and safety associated with chimeric antigen receptor (CAR) T-cell therapy: Results from a comprehensive cancer center. PLoS One 17(12), e0278950 (2022). 10.1371/journal.pone.0278950 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gatwood, J., Masaquel, A., Fox, D., Sheinson, D., James, C., Li, J., et al.: Real-world total cost of care by line of therapy in relapsed/refractory diffuse large B-cell lymphoma. J. Med. Econ. 27(1), 738–745 (2024). 10.1080/13696998.2024.2349472 [DOI] [PubMed] [Google Scholar]
- 15.Costa, S., Scott, D., Steidl, C., Peacock, S., Regier, D.: Real-world costing analysis for diffuse large B-cell lymphoma in British Columbia. Curr. Oncol. 26(2), 108–113 (2019). 10.3390/cancers14040908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Huang, Q., Borra, S., Li, J., Wang, L., Shrestha, S., Sundaram, M., et al.: Time to next Treatment, health care resource Utilization, and costs associated with ibrutinib use among U.S. Veterans with chronic lymphocytic Leukemia/Small lymphocytic lymphoma: A Real-World retrospective analysis. Journal of Managed Care & Specialty Pharmacy 26(10), 1266–1275 (2020). 10.18553/jmcp.2020.20095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cho, J., Jang, S., Kang, D., Lee, E., Koh, H., Yoon, D., et al.: A nationwide analysis of the treatment patterns, survival, and medical costs in Korean patients with relapsed or refractory diffuse large B-cell lymphoma. Front. Oncol. 14, 1282323 (2024). 10.3389/fonc.2024.1282323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Petersen, C., Jensen, K., Rosenzweig, M., von Osmanski, B., Ankarfeldt, M., Petersen, J.: Mapping outcomes and registries used in current Danish pharmacoepidemiological research. Clin. Epidemiol. 14, 521–542 (2022). 10.2147/CLEP.S341480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Erlangsen, A., Fedyszyn, I.: Danish nationwide registers for public health and health-related research. Scand J Public Health 43(4), 333–339 (2015). 10.1177/1403494815575193 [DOI] [PubMed] [Google Scholar]
- 20.Arboe, B., El-Galaly, T., Clausen, M., et al.: The Danish National lymphoma registry: Coverage and data quality. PLoS One 11(6), e0157999 (2016). 10.1371/journal.pone.0157999 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Malignt Lymfom og CLL National årsrapport: 1. januar 2022– 31. december 2022. Dansk lymfom Database/Dansk Kronisk Lymfatisk Leukæmi Database. National årsrapport 2022. (2022). Regionernes Kliniske Kvalitetsudviklingsprogram (www.rkkp.dk).
- 22.Pedersen, C.: The Danish civil registration system. Scand. J. Public. Health 39(7 Suppl), 22–25 (2011). 10.1177/1403494810387965 [DOI] [PubMed] [Google Scholar]
- 23.Lynge, E., Sandegaard, J., Rebolj, M.: The Danish National Patient Register. Scand J Public Health 39(7 Suppl), 30–33 (2011). 10.1177/1403494811401482 [DOI] [PubMed] [Google Scholar]
- 24.Andersen, J., Olivarius, N., Krasnik, A.: The Danish National health service register. Scand. J. Public. Health. 39(7 Suppl), 34–37 (2011). 10.1177/1403494810394718 [DOI] [PubMed] [Google Scholar]
- 25.Kildemoes, H., Sørensen, H., Hallas, J.: The Danish National prescription registry. Scand. J. Public. Health. 39(7 Suppl), 38–41 (2011). 10.1177/1403494810394717 [DOI] [PubMed] [Google Scholar]
- 26.Hjollund, N., Larsen, F., Andersen, J.: Register-based follow-up of social benefits and other transfer payments: Accuracy and degree of completeness in a Danish interdepartmental administrative database compared with a population-based survey. Scand. J. Public. Health. 35, 497–502 (2007). 10.1080/14034940701271882 [DOI] [PubMed] [Google Scholar]
- 27.Cheson, B., Fisher, R., Barrington, S., Cavalli, F., Schwartz, L., Zucca, E., et al.: Recommendations for initial evaluation, staging, and response assessment of hodgkin and non-Hodgkin lymphoma: The Lugano classification. J. Clin. Oncol. 32, 3059–3068 (2014). 10.1200/JCO.2013.54.8800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Coiffier, B., Thieblemont, C., Van Den Neste, E., Lepeu, G., Plantier, I., Castaigne, S., et al.: Long-term outcome of patients in the LNH-98.5 trial, the first randomized study comparing rituximab-CHOP to standard CHOP chemotherapy in DLBCL patients: A study by the groupe d’etudes des lymphomes de l’adulte. Blood 116, 2040–2045 (2010). 10.1182/blood-2010-03-276246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Sehn, L., Gascoyne, R.: Diffuse large B-cell lymphoma: Optimizing outcome in the context of clinical and biologic heterogeneity. Blood 125(1), 22–32 (2015). 10.1182/blood-2014-05-577189 [DOI] [PubMed] [Google Scholar]