Abstract
Background
Patients on kidney replacement therapy (KRT) face high symptom and treatment burden, especially at end of life. Yet insights into end-of-life healthcare utilisation and costs remain limited. We aimed to describe healthcare utilisation and costs during the last year of life across KRT modalities.
Methods
We used Dutch health insurance claims data to identify incident and prevalent patients aged ≥ 65 years, treated with KRT during the year preceding death, and who deceased between June 2016 and December 2021. Healthcare utilisation and costs in the last year of life were analysed for different KRT modalities and compared with controls without kidney disease-related claims, with controls being matched on sex, age, socio-economic status, and year of death.
Results
We identified 7279 patients on KRT (4614 haemodialysis [HD] patients, 766 peritoneal dialysis [PD] patients and 1899 kidney transplant [KTx] recipients) and 14,558 controls. During their last year of life, 85% of patients on KRT experienced ≥ 1 hospital admission, with 27% requiring intensive care - representing to 2.6- and 3.5-fold higher proportions than in their controls, with minimal variation across KRT modalities. Similar differences were observed for emergency department visits, outpatient visits, and long-term care use. Healthcare utilisation significantly increased in the last three months preceding death, more markedly among patients on KRT than controls (p < 0.001). Average healthcare costs were 5.9, 5.5 and 3.0 times higher for patients on HD (€117,520), PD (€108.294), and KTx recipients (€59,489), respectively, compared to controls (€19,820). KRT-unrelated costs, encompassing 40% to 74% of total costs, were 2.2 to 2.6 times higher in patients on KRT compared to controls, with specialist hospital care costs being three times higher.
Conclusion
In the last year of life, healthcare utilisation is higher for patients on KRT compared to controls, with a further increase towards end-of-life, along with higher KRT-unrelated costs.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12882-025-04605-2.
Keywords: Chronic kidney disease, Dialysis, End-of-life costs, Healthcare utilisation, Kidney transplantation
Key learning points
What was known
This study adds
Potential impact
• Kidney replacement therapy (KRT) is amongst the most expensive chronic treatments.
• Patients on KRT frequently also utilise healthcare services for conditions unrelated to KRT.
• Healthcare utilisation and costs typically rise at end of life.
• During the last year of life, patients on KRT were 2.6 times more likely to be admitted to the hospital and 3.5 times more likely to require ICU care compared to patients without kidney disease. This further increased more markedly in the last months of life, with minimal variation across KRT modalities.
• Healthcare costs during the last year of life were respectively 5.9, 5.5 and 3.0 times higher for patients on haemodialysis, peritoneal dialysis and kidney transplant recipients, compared to patients without a kidney disease.
• Patients on KRT incurred three times more costs on KRT-unrelated specialist care.
• Extensive healthcare utilisation places a high treatment burden on patients and caregivers, while also imposing an economic burden on society.
• Insights into healthcare use and costs can raise awareness among healthcare providers about treatment burden and inform advance care planning.
Our findings can facilitate discussion on the allocation of resources to meet patient needs and ensure long-term sustainability of care.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12882-025-04605-2.
Introduction
Patients on kidney replacement therapy (KRT) experience high symptom and treatment burden, which translates into substantial healthcare utilisation [1]. The majority of this healthcare use is related to dialysis or kidney transplantation [2, 3], but costs vary between patients. Among haemodialysis patients, frail patients incur 22% higher healthcare costs than non-frail patients, primarily attributable to increased inpatient care [4]. This difference in healthcare utilisation between frail and non-frail patients further diverged among those approaching the end of life [4]. The last phase of life is accompanied by a decline in health-related quality of life and increased healthcare utilisation [5, 6]. Towards the end of life, treatment goals ideally shift from prolonging life to maximising comfort, particularly when the burden of interventions outweigh the anticipated benefits. For example, in patients with cancer, in-hospital death, intensive care unit admission and chemotherapy in the last week of life were associated with lower quality of life [7]. Additionally, patients with kidney failure are less frequently engaged in advance care planning (ACP) than patients with cancer [8]. Study outcomes reflect intensive end-of-life care among CKD patients, with many patients experiencing unplanned hospital admissions, intensive care unit (ICU) admission, surgeries and mechanical ventilation [6, 9, 10]. However, for patients receiving KRT a detailed description of healthcare utilisation and the associated costs during the last year of life is lacking. Gaining insight in healthcare use is an important initial step in understanding treatment patterns, may lay a foundation for discussion on end-of-life care, and aid policymakers in anticipating future healthcare demands. Therefore, the primary aim of this study was to describe healthcare utilisation and costs during the last year of life of patients aged ≥ 65 year on KRT, stratified by KRT modality, and to compare them with patients without kidney disease. Our secondary aim is to examine changes in healthcare utilisation as patients approach the end of life and to assess whether these changes differ between patients on KRT and their controls. We hypothesized that the high healthcare utilisation further increases in the final months of life, with a more pronounced increase among patients on KRT.
Methods
Data source
A nationwide cohort was defined within the Vektis database, comprising health insurance claims covered by the Dutch Health Insurance Act and Long-Term Care Act, linked with demographic data [11]. The Dutch Health Insurance Act provides mandatory health insurance coverage for medical care (e.g., primary care, hospital care, medication and mental health care). The Long-Term Care Act provides comprehensive long-term care services at home or institutionalised to individuals in need of ongoing assistance with activities of daily living due to chronic illness, disability, or aging. These two acts regulate and cover the vast majority of healthcare services in the Netherlands, where 99% of inhabitants are insured, with the Vektis database encompassing 99% of these insured individuals [12]. Privacy was ensured by pseudonymisation of patients’ national identification number.
Study population
We identified individuals aged ≥ 65 years, who deceased between June 1st, 2016, and December 31st, 2021, with a health insurance claim for dialysis or kidney transplantation during their last year of life. Individuals who initiated HD, PD or KTx more than 365 days before death were categorised as prevalent patients, those who initiated HD, PD or KTx within the year preceding death as incident patients. This classification resulted in six groups. Incident KTx recipients included both pre-emptive transplanted patients and those who underwent dialysis before transplantation (see Supplemental Materials for detailed definition of groups). Further details regarding the proportions of prevalent and incident patients within this cohort, as well as treatment modality switches during the last year of life, were published previously [13]. To prevent the inclusion of individuals with acute kidney injury requiring transient dialysis needs, we excluded incident patients with a dialysis duration of less than 28 days. For prevalent patients, we excluded individuals who were treated with dialysis at 12 months before death but had less than 6 months dialysis during their last year of life. The latter exclusion criterion was applied to account for cases involving transient kidney failure. Analyses were restricted to patients aged ≥ 65 years, who comprise a large fraction of the KRT population and an even larger fraction of decedents, and in whom the balance between life prolongation and quality of life may differ from that in younger patients. After classification, we validated the number of dialysis and KTx patients for each calendar year using an external database, the Dutch Renal Registry [14]. The agreement between the two databases ranged from 90.3% to 100% annually (see Supplemental Materials).
Healthcare utilisation and costs variables
To assess healthcare utilisation, we derived the number of hospital admissions, intensive care unit (ICU) admissions, emergency department (ED) visits, outpatient visits and long-term care outside the hospital. Dialysis treatments were not included in the count of outpatient visits. Variables of healthcare utilisation are based on hospital activity codes and Long-Term Care Act performance codes (listed in Supplemental Materials).
All healthcare costs in the year preceding death were derived and classified for its relation to kidney disease care. KRT-related costs encompassed all expenses related to the dialysis procedure (e.g., access surgery and revisions, dialysis equipment, medication used during dialysis, diagnostics related to dialysis treatment and staff expenses) or kidney transplantation (e.g., transplant surgery, pre- and post-transplant care and donor expenses). KRT-unrelated costs were defined as all remaining costs not directly related to KRT. If we could not definitively attribute expenses to KRT exclusively, we categorised them as KRT-unrelated costs. KRT-unrelated costs were differentiated into the following segments: costs related to primary care, specialist care delivered in the hospital (inpatient and outpatient), mental health care, medication, transport, others (i.e., medical devices, healthcare abroad) and care provided within the Long-term Care Act. In the Netherlands, hospital claims are derived from negotiated administrative prices for high-level groups of diagnoses within a specific time period and cover care provided for a maximum period of 120 days [15]. As such, health insurance claims data provide insights into expenditure, which may be different from actual costs. In this paper, we will use the term costs when we actually refer to expenditure. All costs of a claim are denoted to the first day of that period, irrespective of the actual distribution of care use within the 120-day period. Within the Vektis database, long-term care costs are declared on a date basis. All other costs (i.e., primary care, mental health care, medication, transport and others) are available per calendar year. To account for varying time intervals, we calculated the duration between the index date and the end of that calendar year (i.e., the calendar year before death). We then obtained a proportionate representation of these costs by multiplying the prior year’s costs by the calculated time and adding them to the costs of the calendar year of death. Cost data are reported in euros and adjusted for inflation according to the Dutch Consumer Price index, with 2021 as the base year [16].
Statistical analysis
Patient and treatment characteristics (sex, age at death, KRT duration, KRT switch, and comorbidities) were presented per treatment modality as means with standard deviation for continuous variables and as percentages for categorical variables. We assessed comorbidities using claims of hospital-, pharmaceutical- and primary care, as described elsewhere by Van Oosten et al. [17] and in Supplement 2. Each KRT patient was matched with two controls, which were randomly selected decedents from the Vektis database, provided they had no healthcare claim for CKD or KRT. Controls were matched for sex, age, socio-economic status and year of death. See Supplemental Materials for details on how age and socio-economic status were derived. Records with incomplete demographic data for matching controls were excluded. We assessed covariate balance between patients on KRT and their matched controls using standardized mean differences (SMD). SMD is a summary statistic used for comparison of data, calculated by dividing the mean difference by the pooled standard deviation. Values < 0.1 are considered indicative of a balanced comparison between groups. The chi-square test was performed for comparison of categorical measures and the independent t-test for continuous measures of healthcare utilisation between groups. To compare healthcare utilisation in the quarter preceding death to the prior nine months, we included only prevalent patients for this specific analysis. Since the initiation period of KRT is associated with intensified healthcare utilisation, including incident patients could distort the results based on the timing of their KRT initiation relative to their death. Therefore, for this single analysis, we excluded incident patients. To account for the difference in observation period, healthcare utilisation during the first nine months was averaged per three-month interval, enabling direct comparison with the last three months of life. Costs were presented as means per person per year. Instead of using standard deviations, we presented interquartile ranges as they more accurately reflect the data distribution. Additionally, we calculated the percentage of users that incurred costs.
Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). This article adheres to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines [18].
Results
Baseline characteristics
We enrolled 7279 patients receiving KRT (4614 on HD, 766 on PD and 1899 KTx recipients) and 14,558 matched controls (Fig. 1). Mean age of patients on KRT was 78 ± 7 years, with 63% being male. This was similar for the control group, reflecting successful matching. Comorbidities were more prevalent in patients on KRT compared to controls. The mean KRT duration was 51 ± 31 months, with variation across modalities. KTx recipients had a younger age at death (74 ± 5 years) compared to HD and PD patients (79 ± 7 and 78 ± 6 years respectively). Furthermore, a larger proportion of dialysis patients initiated KRT in the year preceding death (21% of HD and PD patients combined versus 3% of KTx recipients; 16% overall), see Table 1.
Fig. 1.
Flow diagram of participant selection
Table 1.
Patient characteristics per KRT modality and matched controls per treatment group
| Total KRT population (n = 7279) |
Controls (n = 14558) |
SMD* | Per KRT modality | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Haemodialysis | Peritoneal dialysis | Kidney transplantation | |||||||
| Prevalent** (n = 3669) |
Incident (n = 945) | Prevalent** (n = 592) |
Incident (n = 174) | Prevalent** (n = 1843) |
Incident (n = 56***) | ||||
| Male sex, n (%) | 4619 (63%) | 9238 (63%) | < 0.001 | 2,247 (61%) | 629 (67%) | 402 (68%) | 123 (71%) | 1179 (64%) | 39 (70%) |
| Age at death, mean ± SD | 78 ± 7 | 78 ± 7 | 0.006 | 80 ± 7 | 77 ± 6 | 78 ± 6 | 77 ± 6 | 74 ± 5 | 71 ± 4 |
| KRT duration (months), mean ± SD | 51 ± 32 | NA | NA | 53 ± 26 | 8 ± 13 | 38 ± 21 | 7 ± 5 | 77 ± 24 | 37 ± 22 |
| Switch during last year to: | |||||||||
| HD, n (%) | 356 (5%) | NA | NA | - | - | 88 (15%) | 15 (9%) | 252 (13%) | - |
| PD, n (%) | 86 (1%) | NA | NA | 29 (1%) | 17 (2%) | - | - | 50 (3%) | - |
| Comorbidities | |||||||||
| Hypertension, n (%) | 6464 (89%) | 7462 (51%) | 0.898 | 3078 (84%) | 898 (95%) | 557 (94%) | 173 (99%) | 1704 (92%) | 54 (96%) |
| Diabetes mellitus, n (%) | 3536 (49%) | 6320 (25%) | 0.501 | 1654 (45%) | 472 (50%) | 280 (47%) | 82 (47%) | 1013 (55%) | 35 (63%) |
| CAD, n (%) | 1418 (19%) | 692 (5%) | 0.464 | 704 (19%) | 202 (21%) | 139 (23%) | 42 (24%) | 315 (17%) | 16 (29%) |
| CVA/TIA, n (%) | 747 (10%) | 660 (5%) | 0.223 | 368 (10%) | 103 (11%) | 72 (12%) | 17 (10%) | 183 (10%) | 4 (7%) |
| PAD, n (%) | 2076 (29%) | 729 (5%) | 0.509 | 1174 (32%) | 262 (28%) | 177 (30%) | 48 (28%) | 401 (22%) | 14 (25%) |
| Malignancy, n (%) | 2279 (31%) | 4084 (28%) | 0.072 | 962 (26%) | 279 (30%) | 146 (25%) | 49 (28%) | 831 (45%) | 12 (21%) |
*SMD: Standardized mean difference. Controls were matched to cases on sex, age, socio-economic status and calendar year of death
** Prevalent cases were defined as those who initiated KRT > 365 days prior to death, incident cases were defined as those who initiated KRT within the year preceding death provided they received KRT for more than 28 days
*** 37 patients received HD prior to KTx, 10 received PD prior to KTx and < 10 were transplanted pre-emptive. We merged these groups to ensure anonymity
Abbreviations: CAD = coronary artery disease, CVA/TIA = cerebrovascular accident/ transient ischaemic attack, HD = haemodialysis, KRT = kidney replacement therapy, PAD = peripheral artery disease, PD = peritoneal dialysis, SMD = standardized mean difference
Healthcare utilisation in the last year of life
Figure 2 and Table 2 show that patients on KRT were 2.6 times more likely to be admitted to the hospital compared to controls (85% versus 33%), with a median of 14 [4, 29] versus 0 [0, 4] hospital days. The proportion of patients admitted to the ICU was 3.5 times higher for patients on KRT (27% versus 8% in controls. Additionally, patients on KRT were 2.4 and 1.7 times more likely to visit the emergency department and outpatient department, respectively, along with 1.2 times more often long-term care use. Healthcare utilisation was similar across KRT modalities, apart from long-term care. Compared to controls, patients on HD more frequently used long-term care (RR = 1.3, p < 0.01), while patients on PD or KTx showed lower long-term care use (RR = 0.8). In general, incident patients had higher levels of healthcare utilisation compared to prevalent patients, see Supplemental Table S1A-C.
Fig. 2.
Proportion of patients using healthcare facilities during the last year of life in patients on HD, PD and KTx compared to controls. *p < 0.05, ** p ≤ 0.01,*** p ≤ 0.001, NS not significant. Abbreviations: KRT = kidney replacement therapy, ICU = intensive care unit, ED = emergency department
Table 2.
Healthcare utilisation of patients on HD, PD and KTx compared to matched controls during the last year of life
| Total KRT population ( n = 7,279) | Haemodialysis ( n = 4,614) | Peritoneal dialysis ( n = 766) | ||||||||||
| Median [IQR] | Mean ± SD |
Mean difference with controls (95% CI) |
p -value | Median [IQR] | Mean ± SD |
Mean difference with controls (95% CI) |
p -value | Median [IQR] | Mean ± SD |
Mean difference with controls (95% CI) |
p -value | |
| Admission days | 14 [4, 29] | 20.4 ± 22.6 | 16.2 (15.56;16.66) | *** | 13 [4, 28] | 20.0 ± 22.1 | 15.7 (15.06;16.36) | *** | 17 [7, 34] | 23.5 ± 22.1 | 19.2 (17.59;20.86) | *** |
| ICU days | 0 [0, 0] | 1.5 ± 6.0 | 1.1 (0.89;1.19) | *** | 0 [0, 0] | 1.3 ± 6.3 | 0.9 (0.75;1.13) | *** | 0 [0, 0] | 1.2 ± 4.2 | 0.8 (0.53;1.55) | *** |
| No of ED visits | 2 [1, 3] | 2.1 ± 1.9 | 1.5 (1.49;1.58) | *** | 2 [1, 3] | 2.1 ± 1.9 | 1.5 (1.48;1.59) | *** | 2 [0, 3] | 2.3 ± 1.8 | 1.7 (1.61;1.88) | *** |
| No of outpatient visits | 10 [5, 16] | 11.3 ± 8.1 | 7.9 (7.68;8,09) | *** | 9 [5, 14] | 10.2 ± 7.7 | 6.8 (6.61;7.09) | *** | 9 [5, 14] | 10.6 ± 7.1 | 7.2 (6.62;7.68) | *** |
| Days long-term care use | 0 [0, 0] | 40.4 ± 103.8 | 10.5 (7.77;13.28) | *** | 0 [0, 0] | 48.8 ± 113.5 | 18.9 (15.32;22.39) | *** | 0 [0, 0] | 27.3 ± 85.5 | -2.6 (-9.00;3.90) | NS |
| Kidney transplant recipients (n = 1,899) |
Controls ( n = 14, 558) |
|||||||||||
| Median [IQR] | Mean ± SD |
Mean difference with controls (95% CI) |
p-value | Median [IQR] | Mean ± SD | |||||||
| Admission days | 20.4 ± 22.6 | 23.5 ± 22.1 | 19.2 (17.59;20.86) | *** | 0 [0, 4] | 4.3 ± 9.9 | ||||||
| ICU days | 1.5 ± 6.0 | 1.2 ± 4.2 | 0.8 (0.53;1.55) | *** | 0 [0, 0] | 0.4 ± 3.5 | ||||||
| No of ED visits | 2.1 ± 1.9 | 2.3 ± 1.8 | 1.7 (1.61;1.88) | *** | 0 [0, 1] | 0.6 ± 1.1 | ||||||
| No of outpatient visits | 11.3 ± 8.1 | 10.6 ± 7.1 | 7.2 (6.62;7.68) | *** | 1 [0, 5] | 3.4 ± 5.4 | ||||||
| Days long-term care use | 40.4 ± 103.8 | 24.9 ± 80.1 | -5.0 (-8.89;-1.16) | ** | 0 [0, 0] | 29.9 ± 84.9 | ||||||
*Comparison with the controls
*p < 0.05, ** p ≤ 0.01,*** p ≤ 0.001, NS not significant
Abbreviations: ED = emergency department, ICU = intensive care unit, IQR = interquartile range, SD = standard deviation
Healthcare utilisation in the last three months prior to death
Healthcare utilisation intensified in the three months prior to death compared to the preceding nine months in both patients on KRT and their controls. Figure 3 illustrates increases across all groups in hospital and ICU admission days during the last three month, along with increases in the number of ED presentations, outpatient visits and long-term care use. Among prevalent HD patients, the mean number of hospital admission days was 8.4 days in the final three months of life, compared to 2.6 days per quarter in the previous nine months (p < 0.001). In prevalent PD patients and KTx recipients, mean hospital days increased from 3.3 to 12.3 and 3.2 to 10.7 days per quarter, respectively (p < 0.001 for both), see Supplemental Table S2A-C. Patients on KRT exhibited greater increases in healthcare utilisation in the last three months of life compared to controls; the corresponding values are represented in Supplemental Figure S1.
Fig. 3.
Healthcare utilisation rates per quarter of prevalent patients and matched controls, stratified by first nine months (dark colours bars) and the last three months (light coloured bars). *p < 0.05, ** p ≤ 0.01,*** p ≤ 0.001, NS not significant. Abbreviations: ED = emergency department, HD = haemodialysis, ICU = intensive care unit, KTx = kidney transplantation, PD = peritoneal dialysis
Healthcare costs in the last year of life
Mean total healthcare costs of the last year of life were €117,520 for HD, €108,294 for PD and €59,489 for KTx patients. This was higher compared to the controls, i.e., €19,820, see Fig. 4; Table 3. Total healthcare costs in dialysis patients were primarily driven by KRT-related costs, constituting on average €65,216 (55% of total costs) for HD patients and €64,559 (60% of total costs) for PD patients. For kidney transplant recipients, the costs related to KRT varied widely depending on the time since transplantation, with a mean of €3,537 for prevalent patients and €88,491 for those who received a KTx during their last year of life. See Supplemental Table S3 for information on costs for prevalent, incident, switching and non-switching patients per KRT modality.
Fig. 4.
Healthcare costs in the last year of life. Error bars represent the interquartile range. Abbreviations: HD = haemodialysis, KRT = kidney replacement therapy, KTx = kidney transplantation, PD = peritoneal dialysis
Table 3.
Healthcare costs in the last year of life
| Total KRT population (n = 7,279) | Haemodialysis (n = 4,671) | Peritoneal dialysis (n = 709) | Kidney transplant recipients (n = 1,899) | Controls (n = 14,558) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| mean | 25%- 75% | % users | mean | 25%- 75% | % users | mean | 25%- 75% | % users | mean | 25%- 75% | % users | mean | 25%- 75% | % users | |
| Total KRT costs | €52,120 | €18,300 - €72,734 | €65,216 | €63,903- €74,128 | €64,559 | €62,197- €74,020 | €15,262 | €1,140-€10,626 | €0 | €0- €0 | |||||
| Primary care | €1,031 | €364 - €1,262 | 98% | €1,001 | €357- €1,241 | 97% | €1,185 | €445- €1,410 | 99% | €1,046 | €342- €1,256 | 97% | €715 | €13- €942 | 77% |
| Specialist care | €24,486 | €6,550 - €33,452 | 98% | €24,340 | €6,450- €33,082 | 98% | €24,540 | €7,638- €33,551 | 98% | €24,824 | €6,331- €33,882 | 96% | €8,187 | €0- €9,890 | 60% |
| Mental health care | €225 | €0 - €0 | 5% | €224 | €0- €0 | 5% | €191 | €0- €0 | 6% | €238 | €0- €0 | 6% | €226 | €0- €0 | 5% |
| Medication | €4,181 | €2,025 - €5,276 | 98% | €3,811 | €1,805- €4,856 | 97% | €3,818 | €2,030- €4,633 | 99% | €5,226 | €2,716- €6,528 | 98% | €988 | €0- €1,263 | 73% |
| Transport | €4,579 | €1,150 - €6,158 | 92% | €6,062 | €2,603- €7,658 | 99% | €1,561 | €529- €2,171 | 87% | €2,059 | €362- €2,440 | 80% | €458 | €0- €758 | 39% |
| Other | €5,437 | €565 – 5,491 | 94% | €5,621 | €554- €6,014 | 94% | €5,359 | €1,772- €4,982 | 99% | €5,014 | €399- €4,604 | 92% | €1,690 | €0- €1,015 | 61% |
| Long-term care | €9,424 | €0 - €0 | 20% | €11,245 | €0- €7,977 | 13% | € 7,081 | €0- €0 | 13% | €5,821 | €0- €0 | 13% | €7,556 | €0- €0 | 16% |
| Total costs | €101,482 | €74,565-€128,140 | €117,520 | €88,208 - €141,259 | €108,294 | €83,117 - €125,587 | €59,489 | €22,613 - €87,253 | €19,820 | €523 - €26,593 | |||||
Focusing on KRT-unrelated costs, patients on HD incurred 2.6 times those of controls, and PD patients and KTx recipients 2.2 times (Table 3). Specialist care was the main driver of KRT-unrelated costs for all KRT modalities. Compared to controls, patients on KRT had higher costs for long-term care outside the hospital (€9,424 versus €7,556) and three times more costs incurred on specialist care within hospital settings (€24,486 versus €8,187). We specified these KRT-unrelated costs for ten medical specialties with the highest expenditures, see Table 4. Highest costs were related to care of internal medicine, surgery, cardiology and geriatric medicine. Depending on the KRT modality, 69–80% of patients received Internal Medicine care besides KRT-related care, compared to 28% in controls.
Table 4.
Top 10 medical specialties with the highest KRT-unrelated costs
| Haemodialysis | Peritoneal dialysis | Kidney transplantation | Controls | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Medical specialty | Mean costs | % users | Medical specialty | Mean costs | % users | Medical specialty | Mean costs | % users | Medical specialty | Mean costs | % users | |
| 1 | Internal Medicine (apart from KRT-related care) | €7,176 | 76% | Internal Medicine (apart from KRT-related care) | €7,321 | 80% | Internal Medicine (apart from KRT-related care) | €5,639 | 69% | Internal Medicine | €1,550 | 28% |
| 2 | Surgery | €4,182 | 55% | Surgery | €3,358 | 55% | Surgery | €3,073 | 50% | Pulmonology | €1,061 | 20% |
| 3 | Geriatric medicine* | €2,414 | 16% | Cardiology | €3,119 | 71% | Cardiology | €2,174 | 60% | Surgery | €739 | 17% |
| 4 | Cardiology | €2,096 | 65% | Geriatric Medicine* | €2,206 | 14% | Geriatric Medicine* | €2,046 | 14% | Cardiology | €590 | 22% |
| 5 | Neurology | €636 | 35% | Neurology | €795 | 34% | Pulmonology | €1,033 | 31% | Geriatric Medicine* | €518 | 4% |
| 6 | Pulmonology | €609 | 28% | Pulmonology | €548 | 30% | Neurology | €800 | 33% | Neurology | €456 | 16% |
| 7 | Gastroenterology | €456 | 19% | Cardiothoracic Surgery | €446 | 17% | Gastroenterology | €551 | 23% | Radiotherapy | €395 | 7% |
| 8 | Urology | €419 | 20% | Gastroenterology | €390 | 21% | Radiotherapy | €517 | 9% | Gastroenterology | €293 | 9% |
| 9 | Orthopaedics | €293 | 14% | Orthopaedics | €275 | 11% | Cardiothoracic Surgery | €395 | 4% | In-hospital geriatrics* | €257 | 0% |
| 10 | In-hospital geriatrics* | €280 | 23% | Ophthalmology | €264 | 16% | Urology | €386 | 21% | Urology | €187 | 10% |
* In-hospital geriatrics involves care within a hospital setting, whereas geriatric medicine encompasses a wider scope, including long-term care and community-based healthcare
Abbreviations: KRT = kidney replacement therapy
Discussion
In this study we examined healthcare utilisation and costs in the last year of life in a large cohort of deceased patients on KRT. We found significantly higher healthcare utilisation among patients on KRT than among their controls, across all measures and largely irrespective of KRT modality. The increased healthcare utilisation translated into higher healthcare expenditure, with mean KRT-unrelated healthcare expenses in the last year of life more than double those of matched controls without kidney disease. In the final three months of life, patients across all KRT modalities exhibited a greater increase in healthcare utilisation compared to controls.
Consistent with studies involving patients with CKD or kidney failure, we observed extensive levels of healthcare use at end of life in patient on KRT, regardless of the KRT modality [6, 10, 19]. Population-based studies, not focused on kidney disease, have reported a rise in healthcare utilisation as patients near the end of life, particularly in the final months and weeks [20–23]. We found comparable patterns of increasing healthcare utilisation towards death, with a more pronounced rise among patients on KRT compared to controls. Studies in patients with CKD stage 4–5 or on KRT found that 65% preferred to die at home or in a hospice, and an increased likelihood of dialysis withdrawal among those who engage in ACP [24–26], indicating a desire for reduced healthcare intensity at end of life. In our study, depending on the KRT modality, 25–30% of patients were admitted to the ICU at least once in the last year of life. Some other studies in CKD and dialysis patients have indicated higher rates of hospital care use, with ICU admissions ranging from 40 to 49% in the month prior to death [9, 27]. Variations in end-of-life care may be explained by multiple factors in end-of-life settings, whether community-based or hospital-based, including cultural norms towards death, the implementation of ACP, access to palliative care services and healthcare policy. However, fragmentary research on end-of-life healthcare utilisation in patients on KRT limits insights into healthcare system differences. As such, the observations in the present study may differ from end-of-life care provided in other countries, and extrapolation should be approached with caution.
Although several studies have examined healthcare costs in the last year of life [21, 28–30], no research specifically addresses costs in this period for older patients on KRT. Comparing our findings in the last year of life to literature on a typical year showed that KRT-related costs were similar, whereas KRT-unrelated costs were approximately three times higher in the year preceding death [2]. Notably, in our study, we also observed higher costs for KRT-unrelated care in patients on KRT compared to their controls, primarily driven by specialist care within hospital settings. This increased healthcare use may be attributed to a higher prevalence of comorbidities and complications associated with kidney failure. Previous studies have also shown an association between frailty and healthcare utilisation and costs [31, 32].
The main strength of our study lies in its comprehensiveness, with nationwide coverage of health insurance claims data across multiple healthcare domains and a time span of 5.5 years. This enabled us to obtain a thorough overview of costs incurred across all these domains. By using claims data, we avoided response-related selection bias and recall bias. Furthermore, we validated our cohort to the Dutch Renal Registry, ensuring accuracy and completeness of our data.
The primary limitation of our study is the lack of clinical context, which limits our ability to assess whether healthcare utilisation in the last year of life was appropriate or whether clinical decisions were justifiable. Using a health insurance claims database, we were unable to determine whether patients received ACP or how ACP may have influenced healthcare utilisation and costs. This limitation underscores the need to incorporate clinical data in future research to provide a more nuanced understanding of healthcare practices in end-of-life care. We suggest future research to explore variations in end-of-life healthcare utilisation, evaluate how increased healthcare utilisation impacts quality of life and whether enhanced goal setting and ACP could reduce burdensome healthcare utilisation and costs at end-of-life. As a second limitation, our data did not cover two categories of healthcare use that may be used by KRT patients, namely hospice care and home-based care that fall under legislation other than the Long-term Care Act. Palliative care is often embedded within other types of care, and because health insurance claims data lack clinical context, it is not possible to distinguish whether a hospital admission is life-prolonging or palliative in nature. Despite this, we were able to provide robust estimates as the vast majority of healthcare costs were covered. Third, our analysis likely underestimated healthcare costs. Costs are linked to the start date of a claim, which could fall prior to the index date of 365 days before death. Consequently, some costs incurred in the last year of life were not included in our analysis. The non-normal distribution of costs prevented us from imputing these expenses or estimating their magnitude accurately. Furthermore, our study coincided in part with the COVID-19 pandemic, during which routine healthcare services were significantly reduced across all healthcare domains. On the other hand, the COVID-19 pandemic may have contributed to an overestimation of hospital admissions and ICU admissions, as studies have reported increased COVID-19 related admission rates for patients on KRT compared to the general population, along with an increased mortality risk [33–35]. The effects of the COVID-19 pandemic likely applied equally to the control group, minimizing their impact on relative differences. Considering the study duration of 5.5 years, the influence of the COVID-19 pandemic on our results is probably relatively low.
Our findings provide valuable insights for clinicians and policymakers by highlighting the healthcare intensity and economic burden of KRT at the end of life. It is reasonable to assume that healthcare utilisation may entail treatment burden for patients. The intensive healthcare utilisation observed in our study therefore suggests a considerable burden. Accordingly, our results emphasize the importance of exploring care strategies to alleviate this treatment burden. In addition, these insights into the intensity of care can raise awareness among healthcare professionals and inform ACP by supporting discussions between patients and healthcare providers, while also serving as a leverage point in discussions onthe allocation of resources to meet patient needs and ensure the long-term sustainability of care.
In conclusion, our study reveals that patients on KRT have significantly higher healthcare utilisation during their last year of life compared to matched controls, with a further increase towards the end of life. This was accompanied by substantial higher costs, extending beyond KRT-related expenses.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We thank Vektis for providing access to the health claims data. We thank Nefrovisie for data handling and facilitating this analysis/ for support in data management.
Author contributions
The conceptualization of this work was primarily undertaken by C.G.N. Voorend and M. Jongejan. Data analysis was performed by M.J.H Leegte and M. Jongejan. M.J.M van Oosten, M.E. Numans, G.A. de Wit, M. van Buren, A.C. Abrahams, W.J.W. Bos, and C.G.N. Voorend critically reviewed the manuscript. All authors read and approved the final manuscript.
Funding
None declared.
Data availability
The data underlying this study cannot be made publicly available due to privacy regulations and ownership by Vektis. The data is managed by Vektis, and subsets of the database are available for researchers after request and under specific conditions (info@vektis.nl).
Declarations
Ethics approval and consent to participate
This study was conducted using precollected deidentified health insurance claims data obtained from Vektis. Ethical approval was waived by the institutional review board (IRB) of the department Internal Medicine, Leids Universitary Medical Center (LUMC), and need for informed consent was also waived by IRB. The study complied with the ethical principles of the Declaration of Helsinki and applicable legislation on observational studies. In accordance with national regulations and institutional guidelines, ethical approval was not required, as the study did not involve direct patient contact, interventional procedures, or access to identifiable personal data [36]. The study was conducted in compliance with the Declaration of Helsinki. Informed consent was not applicable as all data were entirely unidentifiable.
Consent to publish
Not applicable as individuals are entirely unidentifiable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Fletcher BR, et al. Symptom burden and health-related quality of life in chronic kidney disease: A global systematic review and meta-analysis. PLoS Med. 2022;19(4):e1003954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Mohnen SM, et al. Healthcare costs of patients on different renal replacement modalities - Analysis of Dutch health insurance claims data. PLoS ONE. 2019;14(8):e0220800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.van Oosten MJM, et al. Age-related difference in health care use and costs of patients with chronic kidney disease and matched controls: analysis of Dutch health care claims data. Nephrol Dial Transpl. 2020;35(12):2138–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sy J, et al. The marginal cost of frailty among medicare patients on Hemodialysis. Kidney Int Rep. 2020;5(3):289–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lopes AA, et al. Health-related quality of life and associated outcomes among Hemodialysis patients of different ethnicities in the united states: the Dialysis outcomes and practice patterns study (DOPPS). Am J Kidney Dis. 2003;41(3):605–15. [DOI] [PubMed] [Google Scholar]
- 6.Chesnaye NC, et al. Clinical and patient-reported trajectories at end-of-life in older patients with advanced CKD. Nephrol Dial Transpl. 2023;38(11):2494–502. [DOI] [PubMed] [Google Scholar]
- 7.Zhang B, Nilsson ME, Prigerson HG. Factors important to patients’ quality of life at the end of life. Arch Intern Med. 2012;172(15):1133–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wachterman MW, et al. End-of-Life experience of older adults dying of End-Stage renal disease: A comparison with cancer. J Pain Symptom Manage. 2017;54(6):789–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chen B, et al. Costs of care at the end of life among elderly patients with chronic kidney disease: patterns and predictors in a nationwide cohort study. BMC Nephrol. 2017;18(1):36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chambers S, et al. Health service utilisation during the last year of life: a prospective, longitudinal study of the pathways of patients with chronic kidney disease stages 3–5. BMC Palliat Care. 2018;17(1):57. [DOI] [PMC free article] [PubMed]
- 11.van Oosten MJM, et al. Health claims databases used for kidney research around the world. Clin Kidney J. 2021;14(1):84–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zorgwijzer. Zorgwijzer - Vergelijken is onze zorg. [cited 2023 December 07]; Available from: https://www.zorgwijzer.nl/faq/cijfers-zorgverzekering.
- 13.Jongejan M, et al. Kidney replacement therapy transitions during the year preceding death. Nephrol Dial Transplant; 2024. [DOI] [PMC free article] [PubMed]
- 14.Nefrodata. Dutch Renal Registry. [cited 2024 May 7]; Available from: https://ivisualz.nl/ivisualz/chartFlash/charts.
- 15.Tan SS, et al. DRG systems in europe: variations in cost accounting systems among 12 countries. Eur J Public Health. 2014;24(6):1023–8. [DOI] [PubMed] [Google Scholar]
- 16.Statistics Netherlands (CBS). Consumer prices; price index. [cited 2023 December 11]; Available from: https://www.cbs.nl/en-gb/figures/detail/83131eng.
- 17.van Oosten MJM, et al. Polypharmacy and medication use in patients with chronic kidney disease with and without kidney replacement therapy compared to matched controls. Clin Kidney J. 2021;14(12):2497–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.von Elm E, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med. 2007;4(10):e296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Verberne WR, et al. Hospital utilization and costs in older patients with advanced chronic kidney disease choosing Conservative care or dialysis: A retrospective cohort study. Blood Purif. 2020;49(4):479–89. [DOI] [PubMed] [Google Scholar]
- 20.Langton JM, et al. Retrospective studies of end-of-life resource utilization and costs in cancer care using health administrative data: a systematic review. Palliat Med. 2014;28(10):1167–96. [DOI] [PubMed] [Google Scholar]
- 21.Diernberger K, et al. Healthcare use and costs in the last year of life: a National population data linkage study. BMJ Support Palliat Care; 2021. [DOI] [PubMed]
- 22.Luta X, et al. Healthcare trajectories and costs in the last year of life: a retrospective primary care and hospital analysis. BMJ Support Palliat Care; 2020. [DOI] [PubMed]
- 23.Wong A, et al. Exploring the influence of proximity to death on disease-specific hospital expenditures: a Carpaccio of red herrings. Health Econ. 2011;20(4):379–400. [DOI] [PubMed] [Google Scholar]
- 24.Davison SN. End-of-life care preferences and needs: perceptions of patients with chronic kidney disease. Clin J Am Soc Nephrol. 2010;5(2):195–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kirchhoff KT, et al. Effect of a disease-specific advance care planning intervention on end-of-life care. J Am Geriatr Soc. 2012;60(5):946–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wachterman MW, et al. Association between hospice length of Stay, health care Utilization, and medicare costs at the end of life among patients who received maintenance Hemodialysis. JAMA Intern Med. 2018;178(6):792–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wong SP, Kreuter W, O’Hare AM. Treatment intensity at the end of life in older adults receiving long-term Dialysis. Arch Intern Med. 2012;172(8):661–3. discussion 663-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gielen B, Remacle A, Mertens R. Patterns of health care use and expenditure during the last 6 months of life in belgium: differences between age categories in cancer and non-cancer patients. Health Policy. 2010;97(1):53–61. [DOI] [PubMed] [Google Scholar]
- 29.van der Plas AG, et al. Relating cause of death with place of care and healthcare costs in the last year of life for patients who died from cancer, chronic obstructive pulmonary disease, heart failure and dementia: A descriptive study using registry data. Palliat Med. 2017;31(4):338–45. [DOI] [PubMed] [Google Scholar]
- 30.Polder JJ, Barendregt JJ, van Oers H. Health care costs in the last year of life–the Dutch experience. Soc Sci Med. 2006;63(7):1720–31. [DOI] [PubMed] [Google Scholar]
- 31.Kallenberg MH, et al. Functional and cognitive Impairment, Frailty, and adverse health outcomes in older patients reaching ESRD-A systematic review. Clin J Am Soc Nephrol. 2016;11(9):1624–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hoogendijk EO, et al. Frailty: implications for clinical practice and public health. Lancet. 2019;394(10206):1365–75. [DOI] [PubMed] [Google Scholar]
- 33.Goffin E, et al. COVID-19-related mortality in kidney transplant and haemodialysis patients: a comparative, prospective registry-based study. Nephrol Dial Transpl. 2021;36(11):2094–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Naga YS, et al. Impact of COVID-19 pandemic on care of maintenance Hemodialysis patients: a multicenter study. Clin Exp Nephrol; 2024. [DOI] [PMC free article] [PubMed]
- 35.Jager KJ, et al. Results from the ERA-EDTA registry indicate a high mortality due to COVID-19 in Dialysis patients and kidney transplant recipients across Europe. Kidney Int. 2020;98(6):1540–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.CCMO, C.C.o.R.I.H.S. Additional information on certain types of research - non-WMO research. [cited 2026 March 27]; Available from: https://www.ccmo.nl/onderzoekers/aanvullende-informatie-over-bepaalde-soorten-onderzoek/niet-wmo-onderzoek.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this study cannot be made publicly available due to privacy regulations and ownership by Vektis. The data is managed by Vektis, and subsets of the database are available for researchers after request and under specific conditions (info@vektis.nl).