Abstract
Purpose
Chemotherapy-induced febrile neutropenia (FN) is a life-threatening and chemotherapy dose-limiting adverse event. FN can be prevented with granulocyte-colony stimulating factors (G-CSFs). Guidelines recommend primary G-CSF use for patients receiving either high (> 20%) FN risk (HR) chemotherapy, or intermediate (10–20%) FN risk (IR) chemotherapy if the overall risk with additional patient-related risk factors exceeds 20%. In this study, we applied an EHR text-mining tool for real-world G-CSF treatment evaluation in breast cancer patients.
Methods
Breast cancer patients receiving IR or HR chemotherapy treatments between January 2015 and February 2021 at LUMC, the Netherlands, were included. We retrospectively collected data from EHR with a text-mining tool and assessed G-CSF use, risk factors, and the FN and neutropenia (grades 3–4) and incidence.
Results
A total of 190 female patients were included, who received 77 HR and 113 IR treatments. In 88.3% of the HR regimens, G-CSF was administered; 7.3% of these patients developed FN vs. 33.3% without G-CSF. Although most IR regimen patients had ≥ 2 risk factors, only 4% received G-CSF, of which none developed neutropenia. However, without G-CSF, 11.9% developed FN and 31.2% severe neutropenia.
Conclusions
Our text-mining study shows high G-CSF use among HR regimen patients, and low use among IR regimen patients, although most had ≥ 2 risk factors. Therefore, current practice is not completely in accordance with the guidelines. This shows the need for increased awareness and clarity regarding risk factors. Also, text-mining can effectively be implemented for the evaluation of patient care.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Quantifying healthcare outcomes in clinical practice is a crucial step towards the improvement of cancer patients’ care [1]. The electronic health record (EHR) is a valuable source of real-world medical data, including, e.g., demographics, vital signs, laboratory data, and medication orders, which can be used for treatment evaluation [2, 3]. However, most of the information is stored in unstructured text, specifically regarding treatment outcomes, e.g., in pathology reports, and detailed adverse drug events in narrative notes [4, 5]. Manual data extraction has been the standard extraction method for EHR data, which is known to be time-consuming and error-prone [3, 6]. Novel natural language processing and text-mining techniques facilitate automatized data extraction from EHR [6,7,8] and therefore enable the evaluation of treatments and guidelines in clinical practice.
For years, granulocyte colony-stimulating factors (G-CSF) are used to prevent chemotherapy-induced neutropenia in breast cancer patients [9]. Neutropenia is one of the most serious and common adverse events of myelosuppressive chemotherapy [10, 11]. Complications of neutropenia are fever, or febrile neutropenia (FN), due to opportunistic infections, which often require intravenous antibiotic treatment and hospitalization [12,13,14]. Moreover, patients developing severe FN often receive chemotherapy dose reductions or treatment delays, which is associated with worse survival outcomes [13].
Prophylactic use of granulocyte colony-stimulating factors (G-CSF), e.g., filgrastim, has shown to reduce the severity and duration of neutropenia and the incidence of FN by 50–90% [10, 14, 15]. G-CSF use can also result in adverse events, mainly mild to moderate bone pain (25–36%), but also potentially secondary myeloid neoplasms [16]. Furthermore, broad use of G-CSF was previously assumed to be a significant financial burden to the healthcare system; the cost-effectiveness of G-CSF is highly related with both the FN risk and the G-CSF costs in clinical practice [17, 18].
FN incidence is primarily related to the type and intensity of the chemotherapy regimen. Therefore, European Organisation for Research and Treatment of Cancer (EORTC) and National Comprehensive Cancer Network (NCCN) guidelines indicate that primary prevention of FN with G-CSF (PP G-CSF) should be applied if patients receive treatment with a high-risk (HR) (> 20%) for FN. For patients receiving a chemotherapy regimen with an intermediate FN risk (IR) of 10–20%, other patient-related risk factors should be considered to define if the overall FN risk exceeds 20%, and thus, PP G-CSF is indicated [10, 19, 20]. Although multiple patient-related risk factors are related with an increased FN risk, e.g., age ≥ 65 years, advanced disease, and female gender, the exact risk attribution of these factors to FN is not yet defined [10]. This may complicate decision-making as to whether or not to administer PP G-CSF. It is already shown that in clinical practice, not all patients receive PP G-CSF when recommended [11, 21]. However, real-world evidence on PP G-CSF utilization linked to risk factors for FN is limited.
Even though several guidelines on G-CSF use for clinical practice are present, it is not clear to what extend these are followed. Therefore, the aim of this study is to retrospectively review guideline adherence for HR and IR chemotherapy regimens in a breast cancer population by collecting data from the EHR with text mining.
Methods
Patient population
All patients aged 18 years and older with breast cancer were included if they started a HR or IR chemotherapy regimen between January 2015 and February 2021 at Leiden University Medical Center (LUMC), Leiden, the Netherlands. Patients participating in the DIRECT study were excluded, since these patients by study design were not allowed to receive PP G-CSF treatment and therefore might bias results [22]. The study protocol was reviewed and approved by the Medical Ethics Review Committee of the LUMC, Leiden, who waived the need for informed consent.
Data collection method
We performed the data collection from the EHR with rule-based text-mining software (CTcue B.V., Amsterdam, the Netherlands). This tool enables extraction of structured (e.g., laboratory results and medical prescriptions) and unstructured (free-text notes) data and the immediate conversion of results into a dataset. For unstructured notes, it enables to search for (combinations of) keywords and shows all notes matching with these results. We validated the software previously for data extraction to evaluate metastatic renal cell carcinoma treatments, which showed high accuracy (> 90%) for data collection from structured data [23]. All used queries for patient inclusion and data collection are in supplementary file 1. All used queries for patient inclusion and data collection are available in supplementary file 1. Patients were identified by a combination of the selected chemotherapy treatments in their medication history, mention of the treatment regimen in the notes, and a diagnosis treatment code for breast cancer. All patients and their regimens were manually validated within the software tool. Additionally, for data extracted from unstructured text (tumor receptor characteristics, type of treatment, G-CSF use and incidence of (febrile) neutropenia), we also performed manual validation of the data by reviewing the identified within the tool. G-CSF use and incidence of (febrile) neutropenia were further manually validated by EHR review.
Patient-, disease- and treatment characteristics
All risk factors from the EORTC guideline that could be evaluated in retrospect were included in this study. These were the following patient- and disease characteristics with specified cutoff values per risk factor (RF): age (RF: age > 65 years), sex (RF: female gender), length and weight (RF: body surface area < 2.0 m2), hemoglobin (RF: hemoglobin < 12 g/dl), ALT (RF: abnormal liver transaminases | ALT > 35 U/l), AST (RF: abnormal liver transaminases | AST > 30 U/l), eGFR (RF: renal disease | eGFR < 60 ml/min/1.73m2), absolute neutrophil count (RF: low pretreatment ANC |< 2 × 109 cells/L[24]), serum albumin (RF: albumin < 3.5 g/dl), performance status (RF: performance status > 0), previous treatments (RF: prior chemotherapy), and treatment type (curative or palliative, RF: advanced disease/metastasis) [10].
We did not include the following risk factors: prior episodes of FN, since we estimated the risk prior to cycle one; antibiotic prophylaxis, since patients in The Netherlands do not receive antibiotic prophylaxis; and cardiovascular disease, one or more comorbidities, and prior infections, since the high variability in free-text terminology combined with the uncertainty whether these risk factors are noted structurally in the EHR would result in an incomplete, and potentially incorrect risk estimation.
Outcome measurements
The primary outcome was the incidence of FN from chemotherapy initiation until 21 days after last treatment cycle. FN is defined as the ANC < 0.5 × 109 cells/L, or ANC < 1.0 × 109 cells/L, predicted to fall below 0.5 × 109 cells/L within 48 h, with fever or clinical signs of sepsis. Fever is defined as rise in axillary temperature > 38.5 °C for 1 h [10]. Also, the incidence of grade 3 or higher neutropenia was collected, which is defined as ANC < 1.0 × 109 cells/L (Common Terminology Criteria for Adverse Events v.5.0). We included cases of FN and neutropenia if they met the definition based on structured data, or when noted in unstructured text by a treating physician.
Statistical analysis
Data management and analysis was performed using R 4.1.0 (R CoreTeam, 2021). Descriptive statistics were used to describe patient, treatment, and disease characteristics. The number of patients on HR or IR chemotherapy regimen receiving PP G-CSF that developed neutropenia or FN was summarized in percentages and visualized in a Sankey plot. With the Student’s t-test, the number of risk factors between the subgroups was compared. Also, chi-square tests or Fisher’s exact test if expected frequency was lower than five, was performed to compare risk factors between subgroups.
Results
In total, 190 breast cancer patients, which received an IR or HR regimen between January 2015 and February 2021, were included. All patients were female and had a median age of 52.6 years (± 11.4 years). Baseline patient, tumor, and treatment characteristics are shown in Table 1.
Table 2 presents an overview of the included chemotherapy regimens. In total, patients received 77 HR and 113 IR chemotherapy regimens. Most applied HR regimens were dose dense doxorubicin and cyclophosphamide (ddAC) → paclitaxel and carboplatin (32.5%), ddAC → paclitaxel (31.2%), and the combination of paclitaxel, doxorubicin, and cyclophosphamide (TAC, 24.7%). Most applied IR regimens were a combination of doxorubicin and cyclophosphamide (AC, 52.2%) and AC → docetaxel (25.7%).
Table 3 summarizes the proportion of patients that received PP G-CSF, and developed neutropenia, and FN, stratified per risk group. These results are visualized in Fig. 1. Overall, in 37.9% of chemotherapy regimens, PP G-CSF was administered at the start of the treatment regimen, 88.3% during HR treatments and 3.5% during IR regimens. The incidence of severe (≥ grade 3) neutropenia in the overall cohort was 21.1%; 11.1% of the patients developed FN at least once. The incidence of neutropenia and FN combined was higher in patients treated with IR regimens (41.6%) than in HR-treated patients (18.2%). However, FN incidence in both groups was comparable and around 10%. In the HR treatment group, 33.3% of the patients who did not receive PP G-CSF developed FN, in contrast to 7.3% of the patients who received PP G-CSF. Furthermore, in the IR treatment group, none of the four patients that started with PP G-CSF developed FN and 11.9% in the group who did not receive PP G-CSF.
We analyzed the presence of additional patient-related risk factors in the IR-treated group. A mean of 3.4 risk factors per patient was found, ranging from one to eight risk factors per patient (Fig. 2A). Patients who received PP G-CSF (n = 4) had a mean of 5 risk factors, and without PP G-CSF (n = 109) 3.3 (p = 0.13). No significant difference was found between the distribution of individual risk factors between both groups (Table 4).
Of the patients on an IR regimen that did not receive PP G-CSF, thirteen (11.9%) developed FN (Table 3). Figure 2B shows that these patients had a lower mean of risk factors versus the patients who did not develop FN (2.5 versus 3.4; p = 0.0044). Furthermore, the risk factor BSA below 2 m2 was more prevalent in the group who did not develop FN (Table 4).
Discussion
In this study, we investigated whether EHR text mining could be applied to evaluate G-CSF use among breast cancer patients in clinical practice. The high recall rate both and the results of the within-program manual validation indicate that adherence to guidelines can be performed using EHR text mining. We found that, in general, G-CSFs were not as often administered as primary prophylaxis as indicated in the EORTC guideline. Not all patients (8.8%) who received HR regimens started with G-CSF prophylaxis. Also, even though almost all patients with IR regimens had two or more of the investigated risk factors, only 4% received PP G-CSF. This resulted in an overall neutropenia incidence of 32%, including 11% FN. Our results indicate that prophylactic treatment with G-CSF should be optimized to further prevent the occurrence of (febrile) neutropenia.
High-risk chemotherapy treatment
Despite guidelines indicating 100% of the patients on a HR regimen should receive PP G-CSF, we report a PP G-CSF use of 88%. An undertreatment of the HR population was also found by Gawade et al., who reported 76.4% PP G-CSF use [21]. Also, PP G-CSF use over time seems to be improved over the recent years [25]. In the HR regimen patients that received PP G-CSF according to the guidelines, only 7% still developed FN, which is comparable to the incidence of FN (9.5%) found in breast cancer patients receiving PP G-CSF (intention-to-treat) for 5 days by Clemons et al. [26]. However, approximately a third of the patients without prophylaxis developed FN. This difference may be directly related to the lack of primary FN prophylaxis; however, numbers in this group are small. In our hospital, most of the omissions of PP G-CSF in the HR group were unintentional. This is potentially related to the fact that G-CSFs are not prescribed and ordered as regular medication in the EHR for hospitalized patients. In the future, these errors could be prevented by incorporating G-CSF prescription into the predefined HR chemotherapy treatment protocol in the EHR.
Intermediate-risk chemotherapy treatment
Almost none (4%) of the patients treated with an IR chemotherapy received PP G-CSF. As PP G-CSF should only be administered if the cumulative FN risk per individual patient exceeds 20%, the risk estimation of patient receiving an IR treatment should be based on additional risk factors [10]. The low use of G-CSF in this population is remarkable as all patients have at least one of the investigated risk factors, female gender, and most have more than two, which may indicate standard G-CSF prescription for these patients. In comparison to the 4% of PP G-CSF use in our study, both Gawade et al. and Bacrie et al. reported around 18% of PP G-CSF use in IR risk patients [21, 27]. Even though the low use of G-CSFs in this population, the FN incidence was moderate (12%) and a substantial amount of the patients developed grades 3–4 neutropenia (31%). Higher G-CSF use could further have prevented cases of severe neutropenia, which can also result in chemotherapy dose delay or reductions [28, 29].
None of the investigated risk factors was significantly related to PP G-CSF prescriptions in our patients. It is remarkable that 20 patients (18%) aged > 65 years did not receive PP G-CSF, since this is the most prominent mentioned risk factor in the guidelines [10]. However, none of the risk factors was significantly associated with the incidence of FN in the group of patients who did not receive PP G-CSF. On the contrary, more risk factors were found in the group without FN, and low BSA specifically was more present in this group. Therefore, even though the small sample sizes, it is questionable to what extent the investigated risk factors substantially contribute to FN development and G-CSF prescription in this IR group. To assess whether dose reductions were applied as alternative preventative measure, we assessed dose reductions on the initial dose and found that approximately 5% of patients started at a dose of 70–80% of the standard starting dose. Thus, in our population, dose reductions play only a minor role or not at all in the risk for FN. Also, antibiotic prophylaxis of FN is not used in this population. Therefore, overall, it may have been other, more difficult to assess factors that could have contributed more to both decision-making for PP G-CSF use and real-world FN risk, e.g., risk factors as radiation therapy and the excluded risk factors [12].
Gawade et al., who compared mainly comorbidities to PP G-CSF initiation by retrospectively reviewing a large medical claims database, also showed that risk factors seemed not to influence PP G-CSF initiation in the IR group, although this was suggested in the NCCN guidelines [21]. Furthermore, Lyman et al. compared the model-predicted and physician-predicted FN risk and showed a weak correlation [30]. This underlines that there may be a difference in how physicians weigh patient-, disease- and treatment-related risk factors; therefore, they suggest the need for continuous education on FN risk factors, G-CSF toxicity, guidelines, and appropriate PP G-CSF use. Zooming in, we note that the underlying problem might be the broad definition of some risk factors and their contribution to clinical outcomes. Therefore, besides continuous education, we recommend, firstly, clarification of the risk factors that play a major role in chemotherapy-induced FN, and, secondly, clearer and simple guidelines which state how these risk factors should be weighed. Therefore, the development of a scoring system, comparable to the system developed to estimate the FN risk in patients who receive low-risk treatments, could also be beneficial for the intermediate- and high-risk group [31].
Moreover, G-CSFs were known to be costly and as a consequence, the choice for treatment above > 20% FN risk only was highly related with the cost-effectiveness [15]. However, new biosimilars are proven to be significantly cost-saving and thereby lower the threshold for the application of G-CSF and simplification of the guidelines [32, 33]. This could not only further lower the FN risk, but also the risk on severe neutropenia. Nonetheless, G-CSF can result in adverse events, and potential benefits of G-CSF use should always outweigh risks.
In this study we applied EHR text-mining software. This enabled fast, structured, and pseudonymized patient inclusion and data extraction. Using this method, missing data of the included risk factors, in general, was limited. However, not all potential risk factors could (fully) be assessed retrospectively. Partially, as the EHR is a secondary source and not all patient data are equally well documented in the EHR, e.g., in this study, we could only report the performance status for approximately 60% of the patients. But also, to some extent, because definitions were unclear or too broad, e.g., one or more comorbidities. We still performed manual validation of the critical end-points G-CSF use and FN incidence after text mining, since these were mainly documented in unstructured text. Since only a selection of the data had to be validated, this process was faster compared to complete manual data extraction. Our study shows that a text-mining tool can be an effective method to review adherence to guidelines and that results can be used as a concrete starting point to optimize patient care.
Conclusion
By application of text mining to the EHR, we were able to review G-CSF use in daily practice in breast cancer patients. PP G-CSF use among HR regimen patients was high, however not maximal, and undertreatment resulted in a higher incidence of FN. Most IR regimen patients had more than two risk factors and were therefore entitled to the use of PP G-CSF. However, few received PP G-CSF which could have prevented the occurrence of FN and neutropenia. Therefore, current practice is not completely in accordance with the guidelines, in particular for patients treated with IR regimens, and may result in unnecessary toxicity for patients. We conclude that awareness of risk factors related with neutropenia should be enlarged and these risk factors could be more clearly defined in the guidelines. Finally, our study shows that text-mining methods can be effectively implemented to review daily practice for the evaluation and improvement of patient care.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Code availability
Used queries are summarized in supplementary file 1 and available upon reasonable request.
References
Teisberg E, Wallace S, O’Hara S (2020) Defining and implementing value-based health care: a strategic framework. Acad Med 95(5):682–685
Casey JA, Schwartz BS, Stewart WF, Adler NE (2016) Using electronic health records for population health research: a review of methods and applications. Annu Rev Public Health 37(1):61–81
Cowie MR, Blomster JI, Curtis LH, Duclaux S, Ford I, Fritz F et al (2017) Electronic health records to facilitate clinical research. Clin Res Cardiol 106(1):1–9
Savova GK, Danciu I, Alamudun F, Miller T, Lin C, Bitterman DS et al (2019) Use of natural language processing to extract clinical cancer phenotypes from electronic medical records. Cancer Res 79(21):5463–5470
Liu F, Jagannatha A, Yu H (2019) Towards drug safety surveillance and pharmacovigilance: current progress in detecting medication and adverse drug events from electronic health records. Drug Saf 42(1):95–97
Assale M, Dui LG, Cina A, Seveso A, Cabitza F (2019) The revival of the notes field: leveraging the unstructured content in electronic health records. Front Med (Lausanne) 6:66
Dreisbach C, Koleck TA, Bourne PE, Bakken S (2019) A systematic review of natural language processing and text mining of symptoms from electronic patient-authored text data. Int J Med Inform 125:37–46
Delespierre T, Denormandie P, Bar-Hen A, Josseran L (2017) Empirical advances with text mining of electronic health records. BMC Med Inform Decis Mak 17(1):127
Zardawi SJ, Nordman I, Zdenkowski N (2020) A retrospective analysis of nadir-neutropenia directed pegylated granulocyte-colony stimulating factor on febrile neutropenia rates in (neo)adjuvant breast cancer chemotherapy regimens. Cancer Rep (Hoboken) 3(5):e1266
Aapro MS, Bohlius J, Cameron DA, Lago LD, Donnelly JP, Kearney N et al (2011) 2010 update of EORTC guidelines for the use of granulocyte-colony stimulating factor to reduce the incidence of chemotherapy-induced febrile neutropenia in adult patients with lymphoproliferative disorders and solid tumours. Eur J Cancer 47(1):8–32
Averin A, Silvia A, Lamerato L, Richert-Boe K, Kaur M, Sundaresan D et al (2021) Risk of chemotherapy-induced febrile neutropenia in patients with metastatic cancer not receiving granulocyte colony-stimulating factor prophylaxis in US clinical practice. Support Care Cancer 29(4):2179–2186
Lee M, Yee J, Kim JY, Kim JY, An SH, Lee KE et al (2019) Risk factors for neutropenia and febrile neutropenia following prophylactic pegfilgrastim. Asia Pac J Clin Oncol 15(4):231–237
Lyman GH, Abella E, Pettengell R (2014) Risk factors for febrile neutropenia among patients with cancer receiving chemotherapy: a systematic review. Crit Rev Oncol Hematol 90(3):190–199
Salmon JP, Smakal M, Karanikiotis C, Wojtukiewicz MZ, Omnes Y, DeCosta L et al (2019) Febrile neutropenia (FN) and pegfilgrastim prophylaxis in breast cancer and non-Hodgkin’s lymphoma patients receiving high (> 20%) FN-risk chemotherapy: results from a prospective observational study. Support Care Cancer 27(4):1449–1457
Krol J, Paepke S, Jacobs VR, Paepke D, Euler U, Kiechle M et al (2006) G-CSF in the prevention of febrile neutropenia in chemotherapy in breast cancer patients. Onkologie 29(4):171–178
Lapidari P, Vaz-Luis I, Di Meglio A (2021) Side effects of using granulocyte-colony stimulating factors as prophylaxis of febrile neutropenia in cancer patients: a systematic review. Crit Rev Oncol Hematol 157:103193
Younis T, Rayson D, Jovanovic S, Skedgel C (2016) Cost-effectiveness of febrile neutropenia prevention with primary versus secondary G-CSF prophylaxis for adjuvant chemotherapy in breast cancer: a systematic review. Breast Cancer Res Treat 159(3):425–432
Aarts MJ, Grutters JP, Peters FP, Mandigers CM, Dercksen MW, Stouthard JM et al (2013) Cost effectiveness of primary pegfilgrastim prophylaxis in patients with breast cancer at risk of febrile neutropenia. J Clin Oncol 31(34):4283–4289
Becker PS, Griffiths EA, Alwan LM, Bachiashvili K, Brown A, Cool R et al (2020) NCCN Guidelines insights hematopoietic growth factors version 1. J Natl Compr Canc Netw 18(1):12–22
Smith TJ, Bohlke K, Lyman GH, Carson KR, Crawford J, Cross SJ et al (2015) Recommendations for the use of WBC growth factors: American Society of Clinical Oncology clinical practice guideline update. J Clin Oncol 33(28):3199–3212
Gawade PL, Li S, Henry D, Smith N, Belani R, Kelsh MA et al (2020) Patterns of granulocyte colony–stimulating factor prophylaxis in patients with cancer receiving myelosuppressive chemotherapy. Support Care Cancer 28(9):4413–4424
de Groot S, Lugtenberg RT, Cohen D, Welters MJP, Ehsan I, Vreeswijk MPG et al (2020) Fasting mimicking diet as an adjunct to neoadjuvant chemotherapy for breast cancer in the multicentre randomized phase 2 DIRECT trial. Nat Commun 11(1):3083
van Laar SA, Gombert-Handoko KB, Guchelaar H-J, Zwaveling J (2020) An electronic health record text mining tool to collect real-world drug treatment outcomes: a validation study in patients with metastatic renal cell carcinoma. Clin Pharmacol Ther 108(3):644–652
Jenkins P, Freeman S (2009) Pretreatment haematological laboratory values predict for excessive myelosuppression in patients receiving adjuvant FEC chemotherapy for breast cancer. Ann Oncol 20(1):34–40
Goyal RK, Tzivelekis S, Rothman KJ, Candrilli SD, Kaye JA (2018) Time trends in utilization of G-CSF prophylaxis and risk of febrile neutropenia in a Medicare population receiving adjuvant chemotherapy for early-stage breast cancer. Support Care Cancer 26(2):539–548
Clemons M, Fergusson D, Simos D, Mates M, Robinson A, Califaretti N et al (2020) A multicentre, randomised trial comparing schedules of G-CSF (filgrastim) administration for primary prophylaxis of chemotherapy-induced febrile neutropenia in early stage breast cancer. Ann Oncol 31(7):951–957
Bacrie J, Laurans M, Iorio P, Fourme E, Volters AB, Bozec L et al (2018) Febrile neutropenia in adjuvant and neoadjuvant chemotherapy for breast cancer: a retrospective study in routine clinical practice from a single institution. Support Care Cancer 26(12):4097–4103
Pettengell R, Schwenkglenks M, Leonard R, Bosly A, Paridaens R, Constenla M et al (2008) Neutropenia occurrence and predictors of reduced chemotherapy delivery: results from the INC-EU prospective observational European neutropenia study. Support Care Cancer 16(11):1299–1309
Liutkauskiene S, Grizas S, Jureniene K, Suipyte J, Statnickaite A, Juozaityte E (2018) Retrospective analysis of the impact of anthracycline dose reduction and chemotherapy delays on the outcomes of early breast cancer molecular subtypes. BMC Cancer 18(1):453
Lyman GH, Dale DC, Legg JC, Abella E, Morrow PK, Whittaker S et al (2015) Assessing patients’ risk of febrile neutropenia: is there a correlation between physician-assessed risk and model-predicted risk? Cancer Med 4(8):1153–1160
Klastersky J, Paesmans M, Rubenstein EB, Boyer M, Elting L, Feld R et al (2000) The multinational association for supportive care in cancer risk index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 18(16):3038–3051
McBride A, Campbell K, Bikkina M, MacDonald K, Abraham I, Balu S (2017) Cost-efficiency analyses for the US of biosimilar filgrastim-sndz, reference filgrastim, pegfilgrastim, and pegfilgrastim with on-body injector in the prophylaxis of chemotherapy-induced (febrile) neutropenia. J Med Econ 20(10):1083–1093
McBride A, Wang W, Campbell K, Balu S, MacDonald K, Abraham I (2020) Economic modeling for the US of the cost-efficiency and associated expanded treatment access of conversion to biosimilar pegfilgrastim-bmez from reference pegfilgrastim. J Med Econ 23(8):856–863
Author information
Authors and Affiliations
Contributions
Conceptualization: H. J. Guchelaar, Juliette Zwaveling, Kim B. Gombert-Handoko, Sylvia A. van Laar and Sophie Wassenaar; methodology: Juliette Zwaveling, Kim B. Gombert-Handoko, Sylvia A. van Laar; formal analysis and investigation: Sylvia A. van Laar; writing — initial draft: Sylvia A. van Laar: writing — review & editing: Henk-Jan Guchelaar, Judith R. Kroep, Juliette Zwaveling, Kim B. Gombert-Handoko, Sophie Wassenaar; recourses: Judith R. Kroep; supervision: Henk-Jan Guchelaar, Juliette Zwaveling, Kim B. Gombert-Handoko.
Corresponding author
Ethics declarations
Ethics approval
The study protocol was reviewed and approved by the Medical Ethics Review Committee of the LUMC, Leiden, who waived the need for informed consent.
Consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
van Laar, S.A., Gombert-Handoko, K.B., Wassenaar, S. et al. Real-world evaluation of supportive care using an electronic health record text-mining tool: G-CSF use in breast cancer patients. Support Care Cancer 30, 9181–9189 (2022). https://doi.org/10.1007/s00520-022-07343-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00520-022-07343-5