Long-term health-care utilisation in older patients with cancer and the association with the Geriatric 8 screening tool: a retrospective analysis using linked clinical and population-based data in Belgium

BACKGROUND

Little evidence is available on the long-term health-care utilisation of older patients with cancer and whether this is associated with geriatric screening results. We aimed to evaluate long-term health-care utilisation among older patients after cancer diagnosis and the association with baseline Geriatric 8 (G8) screening results.

METHODS

For this retrospective analysis, we included data from three cohort studies for patients (aged ≥70 years) with a new cancer diagnosis who underwent G8 screening between Oct 19, 2009 and Feb 27, 2015, and who survived more than 3 months after G8 screening. The clinical data were linked to cancer registry and health-care reimbursement data for long-term follow-up. The occurrence of outcomes (inpatient hospital admissions, emergency department visits, use of intensive care, contacts with general practitioner [GP], contacts with a specialist, use of home care, and nursing home admissions) was assessed in the 3 years after G8 screening. We assessed the association between outcomes and baseline G8 score (normal score [>14] or abnormal [≤14]) using adjusted rate ratios (aRRs) calculated from Poisson regression and using cumulative incidence calculated as a time-to-event analysis with the Kaplan-Meier method.

FINDINGS

7556 patients had a new cancer diagnosis, of whom 6391 patients (median age 77 years [IQR 74-82]) met inclusion criteria and were included. 4110 (64·3%) of 6391 patients had an abnormal baseline G8 score (≤14 of 17 points). In the first 3 months after G8 screening, health-care utilisation peaked and then decreased over time, with the exception of GP contacts and home care days, which remained high throughout the 3-year follow-up period. Compared with patients with a normal baseline G8 score, patients with an abnormal baseline G8 score had more hospital admissions (aRR 1·20 [95% CI 1·15-1·25]; p<0·0001), hospital days (1·66 [1·64-1·68]; p<0·0001), emergency department visits (1·42 [1·34-1·52]; p<0·0001), intensive care days (1·49 [1·39-1·60]; p<0·0001), general practitioner contacts (1·19 [1·17-1·20]; p<0·0001), home care days (1·59 [1·58-1·60]; p<0·0001), and nursing home admissions (16·7% vs 3·1%; p<0·0001) in the 3-year follow-up period. At 3 years, of the 2281 patients with a normal baseline G8 score, 1421 (62·3%) continued to live at home independently and 503 (22·0%) had died. Of the 4110 patients with an abnormal baseline G8 score, 1057 (25·7%) continued to live at home independently and 2191 (53·3%) had died.

INTERPRETATION

An abnormal G8 score at cancer diagnosis was associated with increased health-care utilisation in the subsequent 3 years among patients who survived longer than 3 months.

FUNDING

Stand up to Cancer, the Flemish Cancer Society.

End-of-Life Care in the Last Three Months before Death in Older Patients with Cancer in Belgium: A Large Retrospective Cohort Study Using Data Linkage

This study aims to describe end-of-life (EOL) care in older patients with cancer and investigate the association between geriatric assessment (GA) results and specialized palliative care (SPC) use. Older patients with a new cancer diagnosis (2009-2015) originally included in a previous multicentric study were selected if they died before the end of follow-up (2019). At the time of cancer diagnosis, patients underwent geriatric screening with Geriatric 8 (G8) followed by GA in case of a G8 score ≤14/17. These data were linked to the cancer registry and healthcare reimbursement data for follow-up. EOL care was assessed in the last three months before death, and associations were analyzed using logistic regression. A total of 3546 deceased older patients with cancer with a median age of 79 years at diagnosis were included. Breast, colon, and lung cancer were the most common diagnoses. In the last three months of life, 76.3% were hospitalized, 49.1% had an emergency department visit, and 43.5% received SPC. In total, 55.0% died in the hospital (38.5% in a non-palliative care unit and 16.4% in a palliative care unit). In multivariable analyses, functional and cognitive impairment at cancer diagnosis was associated with less SPC. Further research on optimizing EOL healthcare utilization and broadening access to SPC is needed.

Linking clinical and population-based data in older patients with cancer in Belgium: Feasibility and clinical outcomes

INTRODUCTION

Geriatric screening and geriatric assessment (GS/GA) have proven their benefits in the care for older patients with cancer. However, less is known about the predictive value of GS/GA for outcomes. To research this, clinical data on GS/GA can be enriched with population-based data. In this article we describe the methods and feasibility of data linkage, and first clinical outcomes (GS/GA results and overall survival).

MATERIALS AND METHODS

A large cohort study consisting of patients aged ≥70 years with a new cancer diagnosis was established using linked data from clinical and population-based databases. Clinical data were derived from a previous prospective study where older patients with cancer were screened with G8, followed by GA in case of an abnormal result (GS/GA study; 2009-2015). These data were linked to cancer registration data from the Belgian Cancer Registry (BCR), reimbursement data of the health insurance companies (InterMutualistic Agency, IMA), and hospital discharge data (Technical Cell, TCT). Cox regression analyses were conducted to evaluate the prognostic value of the G8 geriatric screening tool.

RESULTS

Of the 8067 eligible patients with a new cancer diagnosis, linkage of data from the GS/GA study and data from the BCR was successful for 93.7%, resulting in a cohort of 7556 patients available for the current analysis. Further linkage with the IMA and TCT database resulted in a cohort of 7314 patients (96.8%). Based on G8 geriatric screening, 67.9% of the patients had a geriatric risk profile. Malnutrition and functional dependence were the most common GA-identified risk factors. An abnormal baseline G8 score (≤14/17) was associated with lower overall survival (adjusted HR [aHR] = 1.62 [1.50-1.75], p < 0.001).

DISCUSSION

Linking clinical and population-based databases for older patients with cancer has shown to be feasible. The GS/GA results at cancer diagnosis demonstrate the vulnerability of this population and the G8 score showed prognostic value for overall survival. The established cohort of almost 8000 patients with long-term follow-up will serve as a basis in the future for detailed analyses on long-term outcomes beyond survival.

Unplanned hospitalizations in older patients with cancer: Occurrence and predictive factors

BACKGROUND

This study aims to investigate the occurrence of unplanned hospitalizations in older patients with cancer and to determine predictive factors.

METHODS

A prospective Belgian multicentre (n = 22), observational cohort study was performed. Patients ≥70 years with a malignant tumor were included. Patients underwent G8 screening followed by geriatric assessment (GA) if abnormal at baseline and were followed for unplanned hospitalizations at approximately three months. Uni- and multivariable regression models were performed to determine predictive factors associated with unplanned hospitalizations in older patients with an abnormal G8.

RESULTS

In total, 7763 patients were included in the current analysis of which 2409 (31%) patients with a normal G8 score and 5354 (69%) with an abnormal G8 score. Patients with an abnormal G8 were hospitalized more frequently than patients with a normal G8 (22.9% versus 12.4%; p < 0.0001). Reasons for unplanned hospitalizations were most frequently cancer related (25.7%) or cancer therapy related (28%). In multivariable analysis, predictive factors for unplanned hospitalizations in older patients with cancer and an abnormal G8 were female gender, absence of surgery, chemotherapy, ADL dependency, malnutrition and presence of comorbidities.

CONCLUSION

Older patients with cancer and an abnormal G8 screening present a higher risk (23%) for unplanned hospitalizations. Predictive factors for these patients were identified and include not only patient and treatment related factors but also GA related factors.

Adherence to geriatric assessment-based recommendations in older patients with cancer: a multicenter prospective cohort study in Belgium

Background

In the general older population, geriatric assessment (GA)-guided treatment plans can improve overall survival, quality of life and functional status (FS). In GA-related research in geriatric oncology, studies mainly focused on geriatric screening and GA but not on geriatric recommendations, interventions and follow-up. The aim of this study was to investigate the adherence to geriatric recommendations and subsequent actions undertaken in older patients with cancer.

Patient and methods

A prospective Belgian multicenter (N = 22) cohort study included patients ≥70 years with a malignant tumor upon oncologic treatment decision. Patients with an abnormal result on the geriatric screening (G8 ≤14/17) underwent GA. Geriatric recommendations were formulated based on GA results. At follow-up the adherence to geriatric recommendations was documented including a description of actions undertaken.

Results

From November 2012 till February 2015, G8 screening was carried out in 8451 patients, of which 5838 patients had an abnormal result. Geriatric recommendations data were available for 5631 patients. Geriatric recommendations were made for 4459 patients. Geriatric interventions data were available for 4167 patients. A total of 12 384 geriatric recommendations were made. At least one different geriatric recommendation was implemented in 2874 patients. A dietician, social worker and geriatrician intervened most frequently for problems detected on the nutritional, social and functional domain. A total of 7569 actions were undertaken for a total of 5725 geriatric interventions, most frequently nutritional support and supplements, extended home care and psychological support.

Conclusions

This large-scale Belgian study focuses on the adherence to geriatric recommendations and subsequent actions undertaken and contributes to the optimal management of older patients with cancer. We identified the domains for which geriatric recommendations are most frequently made and adhered to, and which referrals to other health care workers and facilities are frequently applied in the multidisciplinary approach of older patients with cancer.

Determining clinically important differences in health-related quality of life in older patients with cancer undergoing chemotherapy or surgery

PURPOSE

Using the EORTC Global Health Status (GHS) scale, we aimed to determine minimal clinically important differences (MCID) in health-related quality of life (HRQOL) changes for older cancer patients with a geriatric risk profile, as defined by the geriatric 8 (G8) health screening tool, undergoing treatment. Simultaneously, we assessed baseline patient characteristics prognostic for HRQOL changes.

METHODS

Our analysis included 1424 (G8 ≤ 14) older patients with cancer scheduled to receive chemotherapy (n = 683) or surgery (n = 741). Anchor-based methods, linking the GHS score to clinical indicators, were used to determine MCID between baseline and follow-up at 3 months. A threshold of 0.2 standard deviation (SD) was used to exclude MCID estimates too small for interpretation. Logistic regressions analysed baseline patient characteristics prognostic for HRQOL changes.

RESULTS

The 15-item Geriatric Depression Scale (GDS15), Visual Analogue Scale (VAS) for Fatigue and ECOG Performance Status (PS) were selected as clinical anchors. In the surgery group, MCID estimates for improvement and deterioration were ECOG PS (5*, 11*), GDS15 (5*, 2) and VAS Fatigue (3, 9*). In the chemotherapy group, MCID estimates for improvement and deterioration were ECOG PS (8*, 7*), GDS15 (5, 4) and VAS Fatigue (5, 5*). Estimates with * were > 0.2 SD threshold. Patients experiencing pain or malnutrition (surgery group) or fatigue (chemotherapy group) at baseline showed a significantly stable or improved HRQOL (p < 0.05) after their treatment.

CONCLUSION

The reported MCID for improvement and deterioration depended on the anchor used and treatment received. The estimates can be used to evaluate significant changes in HRQOL and to determine sample sizes in clinical trials.

A Belgian Survey on Geriatric Assessment in Oncology Focusing on Large-Scale Implementation and Related Barriers and Facilitators

OBJECTIVES

The aim of this study is to describe a large-scale, Belgian implementation project about geriatric assessment (=GA) in daily oncology practice and to identify barriers and facilitators for implementing GA in this setting. Design / setting / participants: The principal investigator of every participating hospital (n=22) was invited to complete a newly developed questionnaire with closed- and open-ended questions. The closed-ended questions surveyed how GA was implemented. The open-ended questions identified barriers and facilitators for the implementation of GA in daily oncology practice. Descriptive statistics and conventional content analysis were performed as appropriate.

RESULTS

Qualifying criteria (e.g. disease status and cancer type) for GA varied substantially between hospitals. Thirteen hospitals (59.1%) succeeded to screen more than half of eligible patients. Most hospitals reported that GA data and follow-up data had been collected in almost all screened patients. Implementing geriatric recommendations and formulating new geriatric recommendations at the time of follow-up are important opportunities for improvement. The majority of identified barriers were organizational, with high workload, lack of time or financial/staffing problems as most cited. The most cited facilitators were all related to collaboration.

CONCLUSION

Interventions to improve the implementation of GA in older patients with cancer need to address a wide range of factors, with organization and collaboration as key elements. All stakeholders, seeking to improve the implementation of GA in older patients with cancer, should consider and address the identified barriers and facilitators.

Interaction of CTLA-4 (CD152) with CD80 or CD86 inhibits human T-cell activation

Occupancy of CTLA-4 (cytotoxic T-lymphocyte antigen-4 or CD152) negatively regulates the activation of mouse T lymphocytes, as indicated by the fate of CTLA-4-deficient mice, by the impact of anti-CTLA-4 monoclonal antibodies (mAbs) on mouse T-cell activation in vitro and by the impact of CTLA-4 blockade on the course of experimental tumoral, autoimmune, alloimmune or infectious disease in this animal. The function of human CTLA-4, however, remains less clear. The expression and function of human CTLA-4 were further explored. CTLA-4 was expressed under mitogenic conditions only, its expression being, at least partially, dependent on the secretion of interleukin-2. Memory T cells expressed CTLA-4 with faster kinetics than naive T cells. The functional role of human CTLA-4 was assessed utilizing a panel of four anti-CTLA-4 mAbs that blocked the interaction between CTLA-4 and its ligands. These mAbs, in immobilized form, profoundly inhibited the activation of T cells by immobilized anti-CD3 mAb in the absence of anti-CD28 mAb, but co-stimulated T-cell activation in the presence of anti-CD28 mAb. Finally, and importantly, blockade of the interaction of CTLA-4 with its ligands using soluble anti-CTLA-4 mAbs, in intact form or as Fab fragments, enhanced T-cell activation in several polyclonal or alloantigen-specific CD80- or CD80/CD86-dependent assays, as measured by cytokine production, cellular proliferation or cytotoxic responses. It is concluded that interaction of CTLA-4 with its functional ligands, CD80 or CD86, can down-regulate human T-cell responses, probably by intracellular signalling events and independent of CD28 occupancy.

Human CTLA-4 is expressed in situ on T lymphocytes in germinal centers, in cutaneous graft-versus-host disease, and in Hodgkin's disease

Cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4, CD152) is a molecule expressed on in vitro activated T cells. CTLA-4 shares important sequence homology with CD28 and binds to the same ligands, CD80 (B7-1) and CD86 (B7-2). CTLA-4 probably functions as a negative regulator of T lymphocyte activation in the mouse, although this remains to be proven for human T lymphocytes. We have developed new monoclonal antibodies against human CTLA-4 and have investigated the in situ expression of CTLA-4 in a wide variety of normal and pathological human tissues expressing CD80 and CD86. As revealed in this study, CTLA-4 is expressed on thymocytes in thymic medulla, on a subset of CD4+ T lymphocytes in germinal centers of follicular hyperplasia, on T cells, mainly CD8+, infiltrating skin affected by graft-versus-host disease, and on T cells, mainly CD4+, infiltrating Hodgkin's disease lesions. In immunoelectron microscopy, CTLA-4 was found on the plasma membrane as well as in the hyaloplasm and cytoplasmic vesicles, in agreement with its pattern of expression on in vitro activated T cells. Interestingly, no or at most scarce expression of CTLA-4 was found in granulomatous lymph nodes, T-cell-mediated inflammatory diseases, or non-Hodgkin's lymphomas, regardless of their expression of CD80 or CD86. Thus, expression of CTLA-4 appears to be induced in selective pathological conditions in vivo. The pathways leading to selective induction of CTLA-4 and its role in the pathophysiology of these conditions need to be further investigated.

[Squamous esophageal papillomas]

In three patients, two women of 37 and 58 and a man of 68 years, a papillomatous lesion was incidentally detected in the distal part of the oesophagus. Microscopically a squamous papilloma was seen with parakeratosis and in two patients poikilocytosis, indicating an infection with human papilloma virus (HPV). The lesions were removed endoscopically, in one patient with laser photocoagulation. Squamous papillomas of the oesophagus are benign tumours with a very low incidence. The pathogenesis still remains unclear. Some authors suggest local irritation as a possible mechanism. On the other hand there is growing evidence of an aetiological role of HPV. In several studies the presence of viral antigen or DNA has been demonstrated. These papillomas may undergo malignant transformation, which means that they are best removed preventively.

Lymphocytic gastritis. Clinical and endoscopic presentation and long-term follow-up

Lymphocytic gastritis is a histopathological entity corresponding with diffuse varioliform gastritis but also with other gastroscopic findings. Eighteen patients were followed over a mean period of 25 months. The symptoms, the endoscopic and histopathological abnormalities remained unchanged in the majority of the cases. Conventional peptic ulcer therapy failed to control symptoms or to normalize endoscopic alterations. Helicobacter pylori did not seem to play a role in the pathophysiology. Lymphocytic duodenitis was found in four patients. The relationship between lymphocytic gastritis, Ménétrier's disease and coeliac disease has further to be elucidated.