Health resource utilization associated with skeletal-related events: results from a retrospective European study

The progression of cancer to the skeleton is a common occurrence in patients with advanced disease; at post-mortem examination, 65–90 % of patients with breast or prostate cancer and approximately 35 % of individuals with lung cancer have bone metastases [1, 2] and almost all patients with multiple myeloma develop bone lesions [3]. Metastatic bone disease is the cause of considerable morbidity [2], with affected patients at high risk of experiencing bone complications, also referred to as skeletal-related events (SREs), including radiation to bone, pathologic fracture, surgery to bone, and spinal cord compression [4]. Unless patients are treated with a bone-targeting agent (BTA), SREs may occur as frequently as every 3–6 months [2]. As mobility and functional independence diminish with subsequent SREs, overall health-related quality of life also declines [4]. Furthermore, patients with metastatic bone disease and an SRE have a poorer prognosis and increased risk of death compared with patients who are SRE naïve [5‐7].

Following an SRE, patient care and treatment can be costly, as well as placing a considerable and complex demand on healthcare resources [8‐10]. A retrospective analysis from the Netherlands estimated that the mean per patient cost to treat SREs in individuals with prostate cancer and bone metastases was €6973 (range, €1187–€40,948) [11]. Despite the differences in the healthcare systems in the Netherlands and the UK, similar values have been reported for patients in the UK with breast cancer and bone metastases, with an estimated mean lifetime SRE-associated cost of £11,314–£19,121 (€14,029–€23,710; 1 GBP = 1.24 EUR) [12]. Notably, total medical care costs are substantially higher for patients who have bone metastases and one or more SREs than for those with bone metastases and no SREs (estimated US$48,173 [€37,093; 1 US$ = 0.77 EUR] more per patient per 60 months in the USA) [13].

Although some studies have attempted to estimate SRE-associated costs, there are a limited number of analyses that review specific health resource utilization (HRU) associated with SREs, particularly at a country level within Europe. An analysis from Spain found that patients (N = 28,162) with bone metastases and an SRE required a greater duration of hospital stays and a greater duration of hospital stays due to re-admission, than did patients with bone metastases without SREs [10]. Similarly, patients with bone metastases and SREs had a greater number of hospital readmissions than patients with a primary cancer diagnosis but no metastatic bone disease, suggesting that as the disease progresses, there is a greater HRU burden [10]. In Portugal, a limited retrospective chart review of patients with breast cancer (n = 121) or prostate cancer (n = 31) and at least one SRE occurring within 12 months reported high costs associated with SREs; these costs were predominantly due to hospitalization and medication [8]. In the USA, a prospective study of 238 patients reported substantial HRU associated with SREs, with considerable numbers and durations of inpatient stays, numbers of outpatient visits, and numbers of procedures [9]. The same was concluded from the European cohort of the same study (conducted in Germany, Italy, Spain, and the UK; n = 631) in which all SREs were associated with considerable HRU burden and costs [14, 15].

The availability of country-specific HRU data would help to describe the burden of SREs on individual European healthcare systems and might help when assessing the overall value of new treatment options. These data are of particular relevance given the resource constraints under which many healthcare systems now operate, as they would allow an accurate estimation of resources required to manage patients with SREs and could be used to determine SRE-associated costs for use in budgeting. Thus, this study was conducted to provide country-specific estimates of HRU associated with SREs in eight European countries, where robust data were previously unavailable.

This study is the first multinational, European, before-and-after, retrospective study to describe SRE-associated HRU in real-world practice across a number of tumor types. All SREs were associated with substantial HRU, demonstrated by increases from baseline in the number and duration of inpatient stays, as well as in the number of procedures and outpatient, emergency room, and day-care visits. The primary strength of this study is that the HRU data captured here are representative of clinical practice across eight European countries.

In all countries, all SRE types were associated with increases from baseline in the number and duration of inpatient stays; however, these increases differed according to SRE type. For example, surgery to bone and spinal cord compression contributed up to three times more HRU than radiation to bone. Increases in inpatient HRU for pathologic fractures of long bones were similar to those for surgery to bone or spinal cord compression, and were greater than increases in HRU for fractures affecting other bones, perhaps because longer periods of immobility or more extensive medical interventions were required than for more minor fractures. Our data are consistent with those from other European studies that report substantial inpatient HRU as a result of surgery to bone, spinal cord compression and pathologic fracture [10, 15]; however, our study is the first to distinguish between pathologic fractures of long bones and those of other bones in terms of HRU in clinical practice.

Some variations in the pattern of HRU were observed in different countries, one example being the duration of inpatient stays: Greece and Poland generally had the smallest increases from baseline in duration of inpatient stays, and Switzerland had the largest increases. While small increases in the duration of inpatient stays across most SRE types were seen in Portugal, long-bone pathologic fracture was associated with one of the largest increases in mean duration of inpatient stays of all the countries. This may indicate a difference in the approach to care for this fracture type compared with other SREs in Portugal.

The majority of inter-country differences in HRU were observed in the change from baseline in the numbers of outpatient visits, day-care visits and procedures. For example, in Finland and Portugal, long-bone pathologic fracture accounted for the largest increase from baseline in outpatient HRU of all SREs, whereas in most other countries, this SRE type was associated with small increases in outpatient HRU. This suggests that there are differences in treatment practice, such as the use of less invasive surgical procedures that may account for the increased use of outpatient or day-care facilities in some countries. The increase in the use of outpatient or day-care clinics may also reflect the accessibility of such facilities in certain countries. If patients are required to travel long distances for treatment, this may make overnight stays necessary, thus increasing inpatient stays and concurrently decreasing outpatient visits. In Switzerland, Austria, and the Czech Republic, the largest increases in outpatient visits and number of procedures were for spinal cord compression, but in Greece, radiation to bone accounted for the largest changes. The increases in number of procedures and outpatient and day-care visits associated with radiation to bone in many countries is perhaps unsurprising, as radiotherapy is recommended by guidelines for the treatment of pain associated with metastatic bone disease [16, 17], a symptom experienced by up to 90 % of patients in the later stages of metastatic cancer [16]. A large review of worldwide radiation practice patterns found that multiple fractions of radiation therapy were preferred to single fractions to treat pain associated with bone metastases in Europe [18]. Both multiple-fraction and single-fraction treatment regimens have been shown to be equally effective in palliating pain [19]; therefore, encouraging the use of single-fraction therapy may help to reduce outpatient- and procedure-related HRU.

The HRU associated with SREs reported here, and the subsequent implications for costs [11, 12, 20, 21], highlight the potential reduction in HRU that could be achieved using BTAs to prevent SREs. Indeed, European guidelines recommend BTAs, such as bisphosphonates and denosumab, for patients with bone metastases secondary to advanced malignancies [17, 22‐26]. In patients with breast cancer, ibandronate significantly decreased the incidence of new bone events by 38 % and delayed the time to first SRE compared with placebo (50.6 vs. 33.1 weeks, respectively) [27]. Similar delays in the time to first SRE were observed in patients with breast cancer receiving pamidronate compared with those receiving placebo [28]. In patients with prostate cancer, zoledronic acid significantly delayed the time to first SRE, and reduced the ongoing risk of SREs by 36 %, compared with placebo [29]. A recent study reported that ibandronate was inferior to zoledronic acid for reducing the frequency of SREs in patients with breast cancer [30], and another study reported that zoledronic acid significantly decreased the risk of developing an SRE by an additional 20 % compared with pamidronate [31]. In addition, denosumab was shown to be superior to zoledronic acid in delaying or preventing SREs in patients with advanced solid tumors [32‐36].

Despite the potential benefits of treatment and contrary to guidelines and published evidence, our study found that only 40 % of patients were receiving bisphosphonates at baseline. After experiencing an SRE, 37 % of patients remained untreated and studies suggest that these patients were at risk of further SREs [29, 37] and subsequent resource use. Our data reflect the findings of a large European patient chart audit in which only 53 % of patients with bone metastases received BTA treatment. Furthermore, the audit indicated that 17 % were expected never to receive treatment, reflecting a possible gap in patient care [38].

The patient populations in this study differed from those in clinical trials. Indeed, in clinical trials of BTAs in patients with bone metastases secondary to cancer, asymptomatic SREs are often captured by regular examinations, including as bone scanning, used during follow-up [32, 33, 39]. Recent clinical trials have used an alternative set of endpoints, referred to as symptomatic skeletal-related events (SSEs), comprising radiation to bone, symptomatic pathologic fracture, surgery to bone and symptomatic spinal cord compression [1, 40]. It is likely that all SSEs were captured in the present study, but that some asymptomatic SREs may not have been identified owing to the absence of regular scans.

The main limitation of this study is the probable underestimation of SRE-associated HRU. In this study, most patients had one or two sites of bone metastasis; therefore, the HRU for individuals with multiple bone metastases was probably not fully captured because these patients may require more HRU than those with fewer sites of metastasis. In addition, HRU outside the hospital setting, such as that associated with home visits, is difficult to assess and was not captured in this study. Home visits have been shown to be particularly important in the management of pathologic fractures [9], and may also become important for other SREs, such as spinal cord compression, owing to the loss of mobility and independence that occurs as metastatic bone disease progresses. Furthermore, patients with more advanced disease may be treated in a hospice facility, and hospice HRU may not have been fully captured in our study.

Another limitation of this study is that the distribution of SREs may not accurately reflect that of a real-world setting, owing to recruitment according to predefined targets (in some instances target numbers of patients were not met). In addition, at the Finnish centre, selection bias was observed for radiation to bone. Patients receiving this treatment were enrolled only from palliative wards, where associated HRU may have been higher owing to the level of care delivered at this type of facility. The other HRU data from Finland are, however, considered representative. Despite the limitations, this study provides valuable information on the HRU associated with SREs in clinical practice.

In conclusion, this study demonstrates that in real-world practice, SREs are associated with substantial increases in HRU across all countries investigated. Previous estimates of the contributions of pathologic fractures to HRU may have been imprecise, owing to grouping of all fractures together in other studies. The availability of more effective and better tolerated BTAs to prevent SREs may help to reduce the burden placed on healthcare resources. Further studies on the effect of delaying SREs on HRU and costs in real-world practice are warranted.