Effect of palliative care decisions making on hospital service use at end-of-life in patients with malignant brain tumors: a retrospective study

The real-life data of the brain tumor patient care pathway demonstrate aggressive EOL treatment and poor integration into PC services. Late or no PC decisions, e.g., the decision to terminate curative or life-prolonging anticancer treatments and focus on symptom-centered PC, increased the need for acute services.

Our study population was a deceased brain tumor population treated in a university hospital. In our patient population, 76% had GBM. Overall survival in the present study was relatively short (16 months in all brain tumor patients and 13 months patients with GBM), but in line with the previous studies. Previous retrospective studies [5, 6] have reported similarly short survival rates. In a study of 2,045 GBM patients, Korja et al. [6] reported a median survival of 11.7 months in patients under 70 years old. With patients over 75 years old, the median survival was only 4.5 months. Stupp et al. [5] reported a median survival of 12.1 months with GBM patients who only received radiotherapy and 14.6 months with patients who received combined treatment (chemoradiotherapy and adjuvant temozolomide).

In our study, anticancer treatments (first-line radiotherapy or chemoradiotherapy) were in line with the international guidelines [4‐22]. On the other hand, our patients received slightly more oncological treatments than those in the literature. The majority of our patients (85%) underwent tumor resection or biopsy and received adjuvant treatment. Only 5% of the patients did not receive any treatment. For comparison, in a large retrospective cohort study of 16,717 patients with newly diagnosed GBM, 60% of patients received radiotherapy and chemotherapy or radiotherapy alone, and 33% received no therapy [23]. In another study of 2,836 GBM patients over 70 years of age, 46% of patients underwent surgery and received radiotherapy. 14% of the patients did not undergo either treatment [24]. This difference can be at least partly explained by the lower rate of GBM patients compared to previous studies. Another plausible explanation could be patient selection as in our study only patients who were treated in university hospital were included compared to population based large cohorts.

The same trend was seen in the treatment at progression. One-fourth (25%) of our patients were not treated with second-line therapy at progression, which is somewhat less than in previous reports. A retrospective multicenter analysis of 299 patients with recurrent GBM after first-line treatment reported that 39.5% of the patients did not receive any oncological treatment at progression [25]. Shi et al. [26] reported that 267 out of 637 (42%) patients who progressed after primary treatment for newly diagnosed GBM received neither reirradiation nor systemic treatment at progression. Furthermore, a significant number of the patients in our study (31%) continued oncological treatments during the last 30 days of life. Contrary, in the study of Kuchinad et al. [3], 17 out of 100 GBM patients were treated with chemotherapy in the last 4 weeks of life. This is significantly higher than what we have previously reported from a heterogeneous cancer population of the same region. In all cancer patients, the corresponding figure was 10% [27]. Thus, it seems that the treatment policy was somewhat more aggressive in the present cohort. Interestingly, however, in our cohort the timing of PC decision has a significant impact on aggressiveness of the treatment at end-of-life. Early made PC decision (> 30 days before death) reduced aggressive treatments compared to late decision making. There were similar findings in Harrison’s et al. [28] study, which showed that timing of referral to PC also has an effect on the aggressiveness of the treatment although the PC referral itself did not affect.

PC decision was defined for the majority of the patients (78%) but rather late (median 44 days before death). Very few patients were referred to a PC unit (18%), and this occurred only after the decision to terminate cancer-specific treatment was made, as early integrated PC practice was not used during the study time. Our finding is in line with previous findings that early involvement of PC services is rare in neuro-oncology [20, 29]. One retrospective analysis of 117 deceased GBM patients found that only a third of patients (37%) received PC consultation [13], even though studies have shown that early integrated PC improves quality of life and reduces the use of aggressive treatments at the EOL without impairing survival [30, 31].

Our study indicates that early PC decisions prevent ED visits at the EOL. More than 90% of the patients who visited an ED and were hospitalized in the last 30 days of life did not have a PC decision or it was made very late. Most of those patients who visited an ED (74%) were hospitalized at a university hospital with an average length of stay of 8 days (range 2–16 days). Approximately every second patient (45%) who was hospitalized died during hospitalization. Thus, it seems that an ED visit at the EOL indicates an increased risk of death at a tertiary hospital. Similar findings were found in a single-institution retrospective cohort study, where 37% of GBM patients were hospitalized in the last 4 weeks of life [3]. In addition, Wasilewski et al. [11] reported that 38% of GBM patients had acute care visits within 30 days of death.

In line with our findings, neurological symptoms, most commonly seizures are by far the most common symptom leading to ED and hospitalization among brain tumor patients [11]. More than 50% of patients with GBM experience tumor-related epilepsy [32]. In our study, the most common reasons to visit the ED were neurological symptoms such as mental confusion, seizure and palsy (31%). The next most common reason was deterioration in the patient’s overall clinical health status (24%). Thus, early PC interventions could be particularly valuable in reducing on the use of hospital services at EOL. Investing in PC can have a positive impact on decreasing ED visits and improving documentation on advance care plans [20, 33].

There were some limitations in this study. Our population was quite small and included only patients who had been treated at the Comprehensive Cancer Center of the Helsinki University Hospital. The results do not contain the number of ED visits or inpatient days in primary care services because this information is not in our database. The age of data is another limitation, taking into account the progress that have happened in PC in recent years. At the time of the study, early integrated PC practice was not systematically organized, and it is still not widely available in many countries. Therefore, we believe it is important to show that also timing of PC decision can have an impact on the use of hospital resources. The strength of this study is that we wanted to examine the burden on tertiary hospitals caused by brain tumor patients receiving oncological treatments at EOL. Another strength is the long follow-up period from diagnosis until death.