University of Groningen
Perfect pitstops
Loeffen, Erik
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Loeffen, E. (2019). Perfect pitstops: Towards evidence-based supportive care in children with cancer.
Rijksuniversiteit Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
UNIVERSITY OF GRONINGEN
PERFECT PITSTOPS
Towards evidence-based supportive care in children with cancer
© E.A.H. Loeffen, Groningen 2019
The research described in this thesis was funded by the Alpe d’HuZes foundation / Dutch Cancer Society and a small part was funded by the Netherlands Organisation for Health Research and Development (ZonMW).
The author gratefully acknowledges the financial support for printing this thesis by: Charlie Braveheart Foundation
Cancer Research Center Groningen (CRCG) University Medical Center Groningen (UMCG) University of Groningen (RUG)
Photography: Erik Loeffen
Lay-out and design: Daniëlle Balk | persoonlijkproefschrift.nl Printing: Ridderprint BV | www.ridderprint.nl
ISBN:
978-94-034-1411-9 (printed) 978-94-034-1410-2 (digital)
Perfect Pitstops
Towards evidence-based supportive care in children with cancer
Proefschrift
ter verkrijging van de graad van doctor aan de
Rijksuniversiteit Groningen
op gezag van de
rector magnificus prof. dr. E. Sterken
en volgens besluit van het College voor Promoties.
De openbare verdediging zal plaatsvinden op
woensdag 3 april 2019 om 14:30 uur
door
Erik Arnoldus Henricus Loeffen
geboren op 1 maart 1987
te Wageningen
Promotores
Dr. W.J.E. Tissing
Prof. dr. L.C.M. Kremer
Copromotor
Dr. M.D. van de Wetering
Beoordelingscommissie
Prof. mr. dr. A.A.E. Verhagen
Prof. dr. G.J.L. Kaspers
Prof. dr. J.S. Burgers
Paranimfen
Steven H. Hendriks
Rob A. van Heumen
01
03
05
09
07
11
A
CH AP TE R CH AP TE R CH AP TE R CH AP TE R CH AP TE R CH AP TE R APPE N D IX General introduction Page 12Patients’ and parents’ views regarding supportive care in childhood cancer
Page 52
The importance of evidence-based supportive care practice guidelines in childhood cancer – a plea for their development and implementation
Page 96
Individualized advance care planning in children with life-limiting conditions
Page 206 Reducing procedure-related pain and distress in children with cancer: a clinical practice guideline
Page 148
Treatment-related mortality in children with cancer: prevalence and risk-factors
Page 242
Nederlandse samenvatting, About the author, List of publications, Dankwoord
Page 282
PART III- IMPLEMENTATION
OF CLINICAL PRACTICE
GUIDELINES
02
04
06
10
08
12
CH AP TE R CH AP TE R CH AP TE R CH AP TE R CH AP TE R CH AP TE RDevelopment of clinical practice guidelines for supportive care in childhood cancer – prioritization of topics using a Delphi approach
Page 34
Current variations in childhood cancer supportive care in the Netherlands
Page 72
Reducing pain in children with cancer, a clinical practice guideline: methodology and an overview of the evidence
Page 110
Development of pediatric oncology supportive care indicators: evaluation of febrile neutropenia care in the north of the Netherlands
Page 226 The duration of anthracycline infusion should be at least one hour in children with cancer: a clinical practice guideline
Page 184
General discussion and future perspectives Page 264
PART I - PREPARATION
FOR GUIDELINE
DEVELOPMENT
PART II - DEVELOPMENT OF
CLINICAL PRACTICE
GUIDELINES
PREFACE
Although at first glance it might seem weird to see a Formula 1 pitstop on the cover of a thesis about supportive care in children with cancer, there is actually more than meets the eye. The pitstop is the moment when a racing car stops to change tyres, undergo repairs or mechanical adjustments, or to refuel. In other words, during a pitstop the car is optimally prepared or adjusted to undergo the remaining part of the race. Supportive care in children with cancer is all care except the anti-cancer treatment, comprising things like pain management, antibiotic therapy, and psychological interventions. In other words, with supportive care a child is optimally prepared or treated to undergo the remaining part of his or her treatment. Therefore, in a way supportive care interventions are the pitstops of cancer treatment.
This is not where the analogy ends. In a Formula 1 race, the cars are continuously monitored. When an intervention is deemed necessary, the driver is called into the pitlane. In childhood cancer care, the child is continuously monitored and when an intervention is deemed necessary, the patient is ‘called in’. In addition, in Formula 1 racing, the decision to pit is not solely taken by the engineer(s), but is always discussed with the driver. This is also the case with supportive care, healthcare professionals always discuss the situation and interventions that are deemed appropriate with the child and his/her caregivers.
Then why “Perfect Pitstops”? In the research described in this thesis we have aimed to develop supportive care guidelines, to standardize and optimize care and thus patient outcomes. Formula 1 pitstops are highly protocolized, field-tested operations, that are refined based on previous research and adjusted according to the actual circumstances. This is also true with our guidelines; the recommendations are based on previous study findings and the recommended action should always be evaluated in the light of the patients’ specific circumstances. Thus interventions as perfect as possible.
So therefore “Perfect Pitstops” and the scene of a pitstop on the cover. The child is driving the car, and the symptoms have just been treated (e.g. the ‘pain’ tyre is changed for the ‘no pain’ tyre) by a team of healthcare professionals. The scene is depicted with LEGO® bricks, as are the scenes preceding each chapter. This is done because I like the idea that children who stumble across my thesis (otherwise probably referred to by them as “boring book”) in my consultation room can also enjoy fliping through the pages. It is because all of the above that the choice to have this thesis Formula 1 pitstop themed is the only logical one. The fact that I am a huge Formula 1 fan is of course merely a coincidence.
CHAPTER 1
GENERAL INTRODUCTION
12
CHAPTER 1
1.1 GENERAL INTRODUCTION
“There is no cure for leukemia; treatment is directed at prolonging life and relieving symptoms. […] Without any treatment the average duration of life is approximately four to five months. When transfusions and antibiotics are given, the life expectancy increases to eight months; with a combination of folic acid antagonists, 6-M.P., and steroids given in sequence children live an average of one year.” - Handbook of Pediatrics, Silver et al., 1961.[1]
It is hard to imagine that these lines were published less than 60 years ago. In fact, in the 1961 “Handbook of Pediatrics” the chapter “Neoplastic diseases” only comprised 12 of the total of 576 pages and contained various prognosis paragraphs similar to the one quoted above. Nevertheless, the first signs of the upcoming treatment regimens that would improve survival rates dramatically are already emerging: folic acid antagonists, 6-mercaptopurine and steroids are currently still used in the treatment of acute lymphoblastic leukemia.
In the 60 years since the handbook was published, cure rates have improved to the point that currently four out five children diagnosed with cancer in high-income countries will be cured.[2] Nowadays it is almost hard to imagine that the quote above is about acute lymphoblastic leukemia, as this is one of the types of childhood cancer with the most dramatic improvement in survival rates (see Figure 1.1). This shows how much this field of medicine has changed over the last decades.
Figure 1.1. Overall survival of children diagnosed with acute lymphoblastic leukemia who were enrolled in Children’s Cancer Group and Children’s Oncology Group clinical trials, 1968 to 2009. Source: Hunger et al.[3] Reproduced with permission from New England Journal of Medicine, Copyright Massachusetts Medical Society.
13 GENERAL INTRODUCTION
A major factor in the increase in survival rates is the introduction of multimodal, intensive treatment strategies comprising chemotherapy, radiotherapy, and/or surgery.[2] However these intensive treatments come at a price, as they are associated with several side effects, ranging from mild to potentially life-threatening. Children undergoing anti-cancer therapy suffer from these side effects, which reduce quality of life. In addition, there is a portion of children that die due to these side effects. Supportive care is the care focusing on preventing and treating the side effects of anti-cancer therapy.
There should be a focus on improving supportive care. Because with the ongoing improvement of childhood cancer survival rates, it is not only important if a child survives cancer (although this naturally remains the main concern), but also, and increasingly so, how a child survives cancer. To improve supportive care, and thus reduce treatment-related morbidity and –mortality, development and implementation of evidence-based guidelines is of the utmost importance, as this contributes to informed treatment decisions and uniform care.
1.2 CHILDHOOD CANCER
INCIDENCE AND SURVIVAL
Each year approximately 550 children in the Netherlands are diagnosed with cancer. This is in line with the European incidence rate, which is 14 per 100.000 children (aged 0-14 years).[4] Even though the aforementioned major advances have been made in the treatment of childhood cancer, it remains an important cause of death in children in high-income countries. For instance, in 2015 in the Netherlands, neoplasms were responsible for 43 out of 158 deaths (27%) in children aged one to 10 years, and was therefore the primary cause of death in this age group.[5] In children aged 10-18 years external causes were the primary cause of death, followed by neoplasms with 23% of deaths. For children younger than one year, cancer is only the cause of death in 1% of all deaths, with the majority of deaths being due to conditions originating in the perinatal period and congenital anomalies.
Cancer in children differs to great extent to cancer in adults. For one, a child is growing and developing, which might affect treatment options and choices. In addition, the distribution of cancer types is substantially different in children as compared to adults. [2] Adults mainly suffer from solid malignancies, with carcinomas being the predominant type. In children however, carcinomas only account for 1.5% of all malignancies. Hematological malignancies, with leukemia occurring most frequently, have the highest incidence in childhood cancer, covering nearly half of all diagnosed malignancies in
14
CHAPTER 1
children. Tumors of the central nervous system are also relatively common (23% of all malignancies), as are embryonal tumors (25% of all malignancies). The distribution of the types of cancer is highly dependent of age. For example, retinoblastoma accounts for almost one in 10 cancer diagnoses under the age of one year, but occurs rarely above the age of four years.
All tumor types have different treatment regimens, which are often designed as trial protocols that are adhered to in a national or international manner. In this way, with each protocol update a new treatment approach or adjustment that builds upon the previous findings can be trialed. Treatment protocols can comprise one or more treatment modalities, which are mostly chemotherapy, radiotherapy, surgery and/or immunotherapy. The duration of treatment depends on the type of tumor and various other factors (such as gene variations), and ranges from very short, for example one surgical procedure for some types of germ cell tumors, to very long, for example three years of chemotherapeutical treatment for some types of leukemia.
Although the overall survival in high-income countries has surpassed 80%, there is still great variation in the prognosis of different types of cancer. There are tumor types with an excellent prognosis, such as acute lymphoblastic leukemia (see Figure 1.1). Nevertheless there are also tumors that remain very difficult to treat, such as stage 4 neuroblastoma (10-year overall survival of 38%) or diffuse intrinsic pontine glioma (currently incurable).[6]
TREATMENT-RELATED MORBIDITY AND MORTALITY
Children who undergo treatment for cancer can be confronted with an array of side effects, as each treatment modality has its own side effects profile. Chemotherapy can lead to early toxicities such as nausea and vomiting, neuropathic pain, and/or bone marrow suppression (which on its turn increases the risk of serious infections).[7] But also late effects of chemotherapy occur, such as cardiotoxicity or infertility.[8] Each chemotherapeutic drug has its own profile of side effects, of which the severity can differ from individual to individual. Another widely used treatment modality, radiotherapy, is also associated with both early toxicities (e.g. skin blistering) and late toxicities (e.g. cognitive decline), depending on area and dose of irradiation.[9] Novel treatment modalities are naturally aimed at improving survival, but might also be aimed at decreasing these side effects. For instance, early cognitive outcomes in children irradiated to the brain using proton radiation therapy, a relatively novel treatment modality, show superior results as compared to traditional (photon) radiation therapy.[10]
Morbidity is naturally not limited to the direct somatic toxicities. In fact, when parents were asked what they believed their children found the most bothersome
15 GENERAL INTRODUCTION
cancer treatment-related adverse effects, the most prevalent symptoms were mood swings, fatigue, and disappointment at missing activities with friends/peers.[11] All these treatment-related adverse effects can severely decrease quality of life of a child and his/ her family.
In addition, as the side effects of cancer treatment can be very severe, there is a portion of children that die due to these adverse effects. As the anti-cancer treatments are becoming more and more intensive and the total number of children that die from cancer is decreasing, preventing treatment-related deaths is crucial to further increase survival rates.[12]
1.3 SUPPORTIVE CARE IN CHILDHOOD CANCER
Supportive care is aimed at reducing side effects of childhood cancer treatment in the broadest sense. The Canadian clinician B. Page was among the first to define supportive care in 1994 and did this in both a comprehensive and comprehensible manner: “[..] In other words, supportive care is anything one does for the patient that is not aimed directly at curing his disease but rather is focused at helping the patient and family get through the illness in the best possible condition.”.[13]
Thus, supportive care is an extremely broad field, ranging from topics as medication for reducing nausea and vomiting to distraction techniques for reducing procedural pain, and from providing psychosocial support for a child and his/her family to treating graft-versus-host disease, and so on.
Optimal supportive care is of the utmost importance as it has the potential to reduce or even prevent early and late adverse effects of anti-cancer treatment, hereby increasing quality of life and potentially even reducing mortality.
1.4 EVIDENCE-BASED MEDICINE
In 2007 the BMJ, one of the most important and influential peer-reviewed medical journals, let readers vote on the world’s greatest medical milestones since the mid-19th
century.[14] Sanitation was the clear winner, but at number seven of these most important advances was “evidence-based medicine”.
The term evidence-based medicine (EBM) was introduced to a wide audience in a 1992 publication by Gordon Guyatt from the Canadian McMaster University, still one of the driving forces behind EBM developments today.[15] Guyatt was part of a medical movement that gained momentum in the 1980s, and that argued that medical decisions
16
CHAPTER 1
should be based on the best available scientific knowledge and good arguments. This was inspired by a feeling that medical decisions were too often based on seniority, individual physician habits or successful marketing of pharmaceutical companies.
Since the introduction of EBM, the definition has expanded, stating that healthcare practice should be “based on integrating knowledge gained from the best available research evidence, clinical expertise, and patients’ values and circumstances.”.[16] Several methods to facilitate EBM have been introduced, of which systematic reviews and meta-analyses are among the most important. The body of medical scientific literature is expanding at a speed impossible to keep up with as an individual practitioner. In fact, the MEDLINE database, one of the largest databases with biomedical literature in the world, indexed over 750.000 citations per year in the last 10 years.[17] This corresponds to nearly 90 new medical papers being published every hour. Focusing on childhood cancer, a relatively small clinical area, a simple MEDLINE search for studies published in the last five years using “child AND cancer” resulted in 37.433 citations.[18] This still corresponds to over 20 new childhood cancer studies published per day, an unattainable rate to keep up with for an individual clinician.
Systematic reviews aim to summarize and critically appraise the available scientific knowledge for a clinical question and are therefore a welcome aid for clinicians who practice EBM. In 1993, an international network was formed that aimed to produce high-quality systematic reviews in a collaborative manner.[19] This network was entitled The Cochrane Collaboration, in memory of Archibald Cochrane, a Scottish epidemiologist that is regarded as one of the founders of EBM and an early advocate for the importance of randomized clinical trials. Presently, The Cochrane Collaboration is the largest international organization involved in medical evidence syntheses and its methods are considered as the standard reference for high-quality systematic reviews. Within the field of childhood cancer, Cochrane Childhood Cancer plays an important role in identifying, appraising, and synthesizing the available evidence on relevant clinical questions. The editorial base of Cochrane Childhood Cancer is located in the Princess Máxima Center for Pediatric Oncology in Utrecht, the Netherlands.
Conclusions from a systematic review can be used to develop evidence-based guidelines. These guidelines recommend on care by combining the best available scientific knowledge (systematic review conclusions) with other factors such as a treatment’s acceptability, feasibility, and resource use. When using these guidelines for clinical-decision making, patient relevant factors such as personal wishes and values naturally also play an important role.
17 GENERAL INTRODUCTION
EBM thus stimulated the connection between science and practice by asking structurized clinical questions and developing evidence-based guidelines, and is pivotal to improving care and clinical outcomes. A striking example of this importance is the estimation that if research into preventing and treating AIDS had not been put into practice, currently more than 50% of hospital beds in the US would be filled with AIDS patients.[16] This makes EBM so crucial, as without it new insights would not be implemented in practice and patients would not benefit from the medical scientific advances.
1.5 CLINICAL PRACTICE GUIDELINES
DEFINITION
As there are multiple interpretations of how an evidence-based guideline should be composed, and more importantly when it is considered truly based on evidence, there was a need for a more precise term and definition. In 2011 the Institute of Medicine (IoM) published a document entitled “Clinical Practice Guidelines We Can Trust“, in which the term Clinical Practice Guidelines (CPG) is defined as: “Statements that include recommendations intended to optimize patient care that are informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options.”.[20]
There are various crucial factors in this definition. First; ‘intended to optimize patient care’. Although this seems logical and perhaps even at first sight somewhat condescending, it is important to stress this. Optimizing patient care should always be the underlying main motivator of guideline development. Second; ‘systematic review of evidence’. The CPG should be informed by a thorough and systematic evaluation and appraisal of the existing literature. This comprises searching multiple databases, using a priori developed comprehensive search strategies and search filters, with the results appraised by multiple independent, trained appraisers. Third and final; ‘assessment of the benefits and harms of alternative care options’. Naturally it is important to review the treatment option the CPG is focusing on, but this should always be put in the contemporary context. For example, are there more effective treatments? Are there treatments that are evenly effective but more patient friendly or more cost-effective? Guideline panels should not be fixated at merely studying the treatment they are focusing on, but also assess alternative care options in a systematic and transparent manner.
Although the definition is quite comprehensive, there are various other factors that are considered important parts of CPGs that lack in the IoM definition. Bulkier, but also more comprehensive, the definition of CPGs could be rephrased to: “Statements that
18
CHAPTER 1
include recommendations intended to optimize patient care that are 1) informed by a protocol-driven systematic review of evidence, 2) formulated by a representative panel of professionals and patient representatives, and 3) include an assessment of the benefits, harms, and other relevant outcomes of the care option and alternative care options.”.
EFFECTS OF CPGS
An important effect of CPGs is the reduction of practice variation, as uncertainty about the most effective treatment has been found to be one of the most important causes of practice variations.[21] Variation in practice implies one of the centers is not advising the most effective treatment. Furthermore, being advised a specific treatment in one hospital and a completely other treatment in another hospital can also be confusing and frustrating for patients and their families. Thus in addition to contributing to uniform care, CPGs might also increase the confidence of patients in the provided care.
Reduction of practice variations is important, as it has been shown to be related to better health outcomes. For instance, in a study in children with asthma, variations in practice were associated with worse outcomes as compared to outcomes in children receiving uniform care.[22] With the care for children with cancer in the Netherlands being centralized in 2018 in the new Princess Máxima Center for Pediatric Oncology (Utrecht), care for all these children is coordinated from one center. This creates a unique opportunity to implement CPGs and reduce practice variations (see Figure 1.2).
Figure 1.2. With the Dutch situation changing from seven primary pediatric oncology hospitals to one primary pediatric oncology hospital and several shared care centers, there is a unique oppor-tunity to facilitate uniform care, by going from the ‘classic’ situation (all centers use their internally developed protocols, see left panel) to the ‘desired’ situation (the Princess Máxima Center for Pediatric Oncology endorses supportive care CPGs and disseminates these to the shared care centers, see right panel).
19 GENERAL INTRODUCTION
It has repeatedly been shown that care in line with recommendations from CPGs has improved patient outcomes, in addition to facilitating a more efficient care delivery. [23,24] For instance, multiple studies have shown that guideline-consistent anti-emetic care was associated with a significant reduction of chemotherapy-induced nausea and vomiting in adults.[25,26] The effect of guideline-consistent care can go even further, as it was found that adherence to a CPG in relation to initiating antibiotic therapy in adult low-risk febrile neutropenia was associated with decreased mortality.[27] The effect of guideline adherence on outcomes in supportive care in children with cancer is an area of research that deserves attention but is yet to be explored.
Besides contributing to better health outcomes and uniform care, CPGs have additional benefits. They help clinicians to absorb the rapidly changing scientific state-of-art in a clinical relevant matter, they can improve the knowledge and awareness of patients (through educational materials), they can contribute to changes in public policy, and they expose gaps of evidence for which future studies need to be initiated.[24]
EXISTING GUIDELINES IN SUPPORTIVE CARE IN CHILDHOOD CANCER
Currently the number of CPGs focused on topics in childhood cancer supportive care is limited. Thankfully, for some important topics there are already recent CPGs available. These topics comprise for instance nausea and vomiting, febrile neutropenia, mucositis and fatigue.[7,28-33] Nevertheless, there are numerous important supportive care topics for which evidence-based guidance is lacking. As the field of supportive care is very broad, it is important to first develop CPGs for topics for which there is a high clinical demand for guidance.
An important recent development is the initiation of the International Pediatric Oncology Guidelines in Supportive Care Network (iPOG Network).[34] This is an international, multidisciplinary network of healthcare professionals with an interest in childhood cancer supportive care guideline development. By using international collaboration, the effort that the development of a CPG takes can be shared, and by informing one another on current projects, duplication of work is prevented. In addition, the iPOG network plays a role in educating and supporting individuals or organizations that aim to developed childhood cancer supportive care guidelines.
20
CHAPTER 1
1.6 DEVELOPMENT OF CPGS
Throughout the years, there have been various methods to developing an evidence-based guideline. In an effort to create uniform, sensible, and transparent methods for developing CPGs, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) working group was formed in 2000.[35] This working group specifically focused on developing an approach to grade quality of evidence (or: certainty of effect estimate) and strength of recommendations. Currently, the GRADE approach is the world-leading standard for CPG development. Various organizations also provide local guidance to guideline development (often incorporating the GRADE approach), such as the Dutch National Health Care Institute with the AQUA directory on quality standards.[36]
There are various phases in the development of a CPG, which will briefly be discussed: 1) startup, 2) clinical questions, 3) evidence search and selection, 4) data extraction and quality appraisal, 5) summarizing evidence, 6) formulating recommendations, 7) writing the CPG. An overview of the complete process can be found in Figure 1.3.
In phase one, the startup phase, the topic is selected and a comprehensive guideline development panel (with multiple working groups if applicable) is composed. Also a systematic search is carried out to identify existing, relevant guidelines. If multiple guidelines on the desired topic are identified, the quality should be assessed and concordances and discordances between recommendations should be evaluated. In the case of concordant recommendations, it should be determined if there is sufficient supporting evidence to support these recommendations. For the discordant recommendations (and naturally for topics for which there is a lack of recommendations), clinical questions can be composed.
During phase two, clinical questions are composed by the complete guideline development panel. These should be composed using the PICO structure, an acronym that stands for Patient (who is your target population?), Intervention (which treatment do you want to study?), Control (which treatment do you want to compare it to?), and Outcomes (on which benefits and harms do you want information?). It is important to decide beforehand on the hierarchy of outcomes. In accordance to GRADE, outcomes can be critical for decision making, important but not critical for decision making, or not important for decision making.[37] This is scored by all panel members on a 9-point scale. Deciding this hierarchy a priori will facilitate the process of formulation of recommendations later on in the process.
21 GENERAL INTRODUCTION
When the PICO(s) are finalized, electronic search strategies can be designed. It is becoming increasingly popular to have these search strategies peer reviewed for correctness and applicability, e.g. by a librarian or a medical information specialist. In addition, at this point the eligibility criteria for the study should be determined.
Hereafter, in phase three, the literature search is performed and the selection of relevant citations can commence. This should be done by at least two independent reviewers, whereby discrepancies should be discussed and solved by consensus or by use of a third party arbiter. Usually, this selection is performed in multiple rounds depending on number of citations, for instance first selection based upon titles and abstracts, then based on full-text articles. It is advised to pilot the selection process to identify sub-optimally defined criteria or other chances for optimization.
Phase four is performed in a similar dual, independent fashion as phase 3. In this phase the data is extracted from the relevant studies, and the studies are evaluated for their methodological quality. For the data extraction a purpose-build data extraction form should be used (and piloted), that facilitates unambiguous extraction of relevant data. The quality appraisal should be performed by using an appropriate and established instrument for the specific study type. A well-known example is the Cochrane Risk of Bias Tool for evaluating the quality of randomized controlled trials (RCTs).
When the study qualities are determined, the quality of the body of evidence per outcome can be summarized using the GRADE system.[35] For this purpose all the included studies per outcome are combined and the quality of the body of evidence is graded as high, moderate, low, or very low. The classification is dependent upon multiple factors: the initial study design (e.g. RCTs start as high quality, observational studies as low quality), factors that require upgrading (dose response effect, large magnitude of effect), and factors that require downgrading (study limitations, inconsistency, indirectness, imprecision, or publication bias).
In phase five all findings are summarized in Summary of Findings tables, including conclusions of evidence with an overall level of evidence (corresponding to the lowest level of evidence of included critical outcomes). The Summary of Findings tables facilitate guideline development panel members to have a structured and comprehensive overview of the information needed to base recommendations upon. These tables are discussed and finalized in accordance with the entire panel.
In phase six the guideline development panel will discuss the evidence and formulate recommendations. These recommendations can be classified as strong or weak and can be for or against an intervention. Strong recommendations, phrased as “we recommend…”, imply that most informed patients would want to have that treatment,
22 CHAPTER 1
and most clinicians would want to prescribe that treatment. Weak recommendations, phrased as “we suggest…”, imply that a large portion of informed patients would want that treatment, but also many would not, and puts more focus on identifying individual patient needs to base a joint decision upon.
In 2016 the GRADE working group introduced Evidence-to-decision (EtD) tables, which facilitate the process of discussing the important factors for basing a recommendation upon.[38,39] The EtD framework comprises 11 questions in six domains, i.e. problem, benefits and harms, resource use, equity, acceptability, and feasibility. After discussing and answering the questions, the panel then decides on the balance of desirable and undesirable consequences on a 5-point scale. After this a recommendation strength is chosen and the recommendation is formulated. This is supplemented with a short justification, subgroup considerations, implementation considerations, monitoring and evaluation considerations, and future research priorities. This systematic and transparent approach to formulating recommendations not only helps the guideline development panel, it also makes their choices and considerations insightful for users of the CPG.
When the formulations are finalized the CPG can be written (phase seven). The final document should be reviewed and agreed upon by all guideline development panel members, and can then be send out for external review by stakeholders.
IMPLEMENTATION
As bringing a CPG development to a successful end is a lot of work, it might be tempting to then breathe a sigh of relief and shift focus to another project. This would however be a mistake, as the recommendations of a finalized CPG still have to be made aware to clinicians and patients. It has proven to be challenging to implement a CPG successfully.
23 GENERAL INTRODUCTION
Figure 1.3. An overview of the CPG development process, following GRADE methodology. Source: GRADE Handbook.[40]
Various factors have been identified that are potential barriers to successful guideline implementation, for instance unfamiliarity with guidelines, reluctance of physicians to change their approach, and practical barriers (e.g. availability of recommended treatment).[41,42]
Multiple studies have evaluated the effectiveness of implementation strategies, which proved to be a difficult concept to measure.[41,43] Although there is no gold standard yet for implementation, there are various interventions identified that increase the chance of a successful CPG implementation. Besides interventions specifically tailored to the guideline recommendations (e.g. making the recommended treatment more easily available), these comprised use of computerized decision-making systems, formal training, active engagement of clinicians, and, perhaps most importantly, use of multifaceted implementation strategies combining the aforementioned interventions. [41,43,44]
A method to evaluate the implementation of guidelines, is to use indicators. Indicators are measurable items that can evaluate structure of care (settings in which care occurs), process of care (processes that belong to giving and receiving care), and
24
CHAPTER 1
outcome of care (states of health or events that follow care).[45] The aim of developing and implementing indicators is to monitor, evaluate, and improve the quality of patient care and care deliverance.
Recently, interest in indicators is growing. Several sets of indicators to evaluate quality of care in multiple subdomains of adult oncology have been published and implemented. [46-48] Also the effect of implementing indicators to improve quality of care has been studied, for which it was found that feedback reports are among the most effective implementation strategies.[49] In childhood cancer supportive care, development and implementation of indicators is an underexposed area of research that should be explored, as it has the potential to improve quality of care.
1.7 AIMS AND OUTLINE OF THIS THESIS
The aim of the research described in this thesis is to improve childhood cancer supportive care, by developing and implementing supportive care CPGs for topics for which there is a high need for guidance. The research can be divided in three parts: preparation, development, and implementation.
PART I. PREPARATION FOR GUIDELINE DEVELOPMENT
Firstly, we wanted to explore for which supportive care topics there was a high clinical need for guidance. Thus in Chapter 2 we asked a multidisciplinary group of professionals
to prioritize topics for guideline development. This was done in a two-round Delphi survey were side effects were first scored for prevalence, severity, and adequate treatment options, and subsequently for importance to develop a CPG.
We also wanted to explore the views of patients and parents on supportive care, use of guidelines, and shared decision making. Therefore, in Chapter 3 we report on
a qualitative study in which we held traditional focus group meetings with parents and online asynchronous focus group meetings with children aged 12-18 years.
In Chapter 4 we explored the current state of supportive care practice in the
Netherlands, focusing on concordances and discordances between supportive care practice in the then primary pediatric oncology hospitals. We also evaluated if current practice was in line with recommendations from existing supportive care CPGs.
The preparation phase culminated in a plea (Chapter 5) for the development and
implementation of CPGs for childhood cancer supportive care, in which we argue why these guidelines are desperately needed.
25 GENERAL INTRODUCTION
PART II. DEVELOPMENT OF CLINICAL PRACTICE GUIDELINES
In Chapter 6 we present the methodology for the development of a CPG we developed
from scratch, focused on assessment and management of pain in children with cancer. In this chapter also the overall results of the inclusion and appraisal of evidence are discussed. As this CPG is very extensive, encompassing 22 clinical questions that we aimed to answer, we opted to publish this guideline in three parts. In Chapter 7, one
of these parts is presented, i.e. the part that focused on the pharmacological, physical, and psychological treatment of procedure-related pain and distress in children with cancer. The other two parts focused on assessment of pain in children with cancer, and pharmacological, physical, and psychological treatment of tumor- and toxicity-related pain in children with cancer.
Another approach to develop a CPG is to use an existing high-quality evidence synthesis and update its findings, whereafter these can be discussed by the guideline development panel and recommendations can be formulated. In Chapter 8 we present
such an approach, that we used to develop a CPG in which we provided recommendations on the infusion duration of anthracycline chemotherapy agents in children with cancer.
PART III. IMPLEMENTATION OF CLINICAL PRACTICE GUIDELINES
To evaluate how guidelines can be implemented we explored various approaches. In
Chapter 9 we aimed to put an existing CPG for pediatric palliative care more firmly
into practice, as it was previously proved to be underused. We did this by developing and piloting a functional individualized pediatric palliative care plan that covers physical, psychological, spiritual and social functioning, and puts great emphasis on the recommendations from the CPG, advance care planning and patients’ and parents’ preferences and desires.
In Chapter 10 we explored how we could develop and use indicators to measure
the current state of care, more specifically to evaluate the implementation of a nationally endorsed febrile neutropenia guideline, and to produce baseline measurements on local quality of febrile neutropenia care.
Not all topics lend itself for indicator development, as sometimes there is not sufficient information available to define these indicators upon. Therefore, in Chapter 11
we tried to define grounds for indicators for treatment-related mortality by exploring in a cohort of over 1750 children with cancer how many children died, how many deaths were treatment-related, and if we could identify risk factors for treatment-related death. In Chapter 12 the findings of all studies are summarized and discussed, and future
perspectives are presented.
26
CHAPTER 1
1.8 REFERENCES
[1] Silver HK, Kemp CH, Bruyn HB. Handbook of Pediatrics. 4th edition. Los Altos, California, United States of America: Lange Medical Publications; 1961. p301-302. [2] Kaatsch P. Epidemiology of childhood cancer. Cancer Treat Rev. 2010 Jun;36(4):277–85. [3] Hunger SP, Mullighan CG. Acute Lymphoblastic Leukemia in Children. Longo DL,
editor. N Engl J Med. 2015 Oct 15;373(16):1541–52.
[4] Steliarova-Foucher E, Stiller C, Kaatsch P, Berrino F, Coebergh J-W, Lacour B, et al. Geographical patterns and time trends of cancer incidence and survival among children and adolescents in Europe since the 1970s (the ACCIS project): an epidemiological study. Lancet. 2004 Dec;364(9451):2097–105.
[5] Statistics Netherlands. CBS Statline - Deaths; underlying cause of death [Internet] . 2018 [cited 2018 Oct 8] . Available from: http://statline.cbs.nl/ Statweb/publication/?DM=SLEN&PA=705 2eng&D1=a&D2=0&D3=1-4&D4=65& LA=EN&HDR=G1,G2,G3&STB=T&VW=T
[6] Berthold F, Spix C, Kaatsch P, Lampert F. Incidence, Survival, and Treatment of Localized and Metastatic Neuroblastoma in Germany 1979–2015. Pediatr Drugs. 2017;19(6):577–93.
[7] Dupuis LL, Boodhan S, Sung L, Portwine C, Hain R, McCarthy P, et al. Guideline for the classification of the acute emetogenic potential of antineoplastic medication in pediatric cancer patients. Pediatr Blood Cancer. 2011 Aug;57(2):191–8.
[8] van Dalen EC, van der Pal HJH, Kremer LCM. Different dosage schedules for reducing cardiotoxicity in people with cancer receiving anthracycline chemotherapy. Cochrane database Syst Rev. 2016;3:CD005008.
[9] Duffner PK. Long-Term Effects of Radiation Therapy on Cognitive and Endocrine Function in Children With Leukemia and Brain Tumors. Neurologist. 2004;10(6):293–310. [10] Gondi V, Yock TI, Mehta MP. Proton therapy for paediatric CNS tumours-improving
treatment-related outcomes. Nat Rev Neurol. 2016;12(6):334–45.
[11] Dupuis LL, Milne-Wren C, Cassidy M, Barrera M, Portwine C, Johnston DL, et al. Symptom assessment in children receiving cancer therapy: the parents’ perspective. Support Care Cancer. 2010;18(3):281–99.
[12] Ethier MC, Blanco E, Lehrnbecher T, Sung L. Lack of clarity in the definition of treatment-related mortality: Pediatric acute leukemia and adult acute promyelocytic leukemia as examples. Blood. 2011;118(19):5080–3.
[13] Page B. What is supportive care? Can Oncol Nurs J. 1994;4:62–3.
[14] Kamerow D. Milestones, tombstones, and sex education. BMJ. 2007 Jan 20;334(7585):126.
27 GENERAL INTRODUCTION
[15] Guyatt G. Evidence-Based Medicine. JAMA. 1992;268(17):2420.
[16] Dickersin K, Straus SE, Bero LA. Evidence based medicine: increasing, not dictating, choice. BMJ. 2007 Jan 6;334(suppl_1):s10–s10.
[17] MEDLINE® Citation Counts by Year of Publication (as of mid - December 2017) [Internet]. U.S. National Library of Medicine; [cited 2018 Jul 27]. Available from: https://www.nlm.nih.gov/bsd/medline_cit_counts_yr_pub.html
[18] Health NI of. PubMed.gov search “child AND cancer” AND last 5 years [Internet]. US National Library of Medicine. [cited 2018 Aug 23]. Available from: https://www.ncbi. nlm.nih.gov/ pubmed/?term=child+AND+cancer+AND+%22last+5+years%22[pdat] [19] Higgins JPT GS (editors). Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0 [Internet]. The Cochrane Collaboration; 2011. Available from: handbook.cochrane.org.
[20] Graham R, Mancher M, Wolman DM. Clinical Practice Guidelines We Can Trust. The National Academies Press, Washington D.C., United States of America; 2011. [21] Greenfield S. Variations in Resource Utilization Among Medical Specialties and
Systems of Care. JAMA. 1992;267(12):1624.
[22] Garbutt JM, Yan Y, Strunk RC. Practice Variation in Management of Childhood Asthma Is Associated with Outcome Differences. J allergy Clin Immunol Pract. 2016;4(3):474–80.
[23] Grimshaw J, Russell I. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet. 1993;1317–22.
[24] Lugtenberg M, Burgers JS, Westert GP. Effects of evidence-based clinical practice guidelines on quality of care: a systematic review. Qual Saf Health Care. 2009 Oct;18(5):385–92.
[25] Aapro M, Molassiotis A., Dicato M, Pelaez I, Rodriguez-Lescure A., Pastorelli D, et al. The effect of guideline-consistent antiemetic therapy on chemotherapy-induced nausea and vomiting (CINV): the Pan European Emesis Registry (PEER). Ann Oncol. 2012;23(8):1986–92.
[26] Gilmore BJW, Peacock NW, Gu A, Szabo S, Rammage M, Sharpe J, et al. Antiemetic Guideline Consistency and Incidence of Chemotherapy-Induced Nausea and Vomiting in US Community Oncology Practice : INSPIRE Study. J Oncol Pract. 2014 Jan;10(1):68-74.
[27] Wright JD, Neugut AI, Ananth C V., Lewin SN, Wilde ET, Lu Y-S, et al. Deviations From Guideline-Based Therapy for Febrile Neutropenia in Cancer Patients and Their Effect on Outcomes. JAMA Intern Med. 2013;173(7):559.
28
CHAPTER 1
[28] Dupuis LL, Boodhan S, Holdsworth M, Robinson PD, Hain R, Portwine C, et al. Guideline for the Prevention of Acute Nausea and Vomiting Due to Antineoplastic Medication in Pediatric Cancer Patients. Pediatr Blood Cancer. 2013;60(January):1073–82.
[29] Flank J, Robinson PD, Holdsworth M, Phillips R, Portwine C, Gibson P, et al. Guideline for the Treatment of Breakthrough and the Prevention of Refractory Chemotherapy-Induced Nausea and Vomiting in Children With Cancer. Pediatr Blood Cancer. 2016 Jul;63(7):1144–51.
[30] Dupuis LL, Robinson PD, Boodhan S, Holdsworth M, Portwine C, Gibson P, et al. Guideline for the prevention and treatment of anticipatory nausea and vomiting due to chemotherapy in pediatric cancer patients. Pediatr Blood Cancer. 2014 Aug;61(8):1506–12.
[31] Sung L, Robinson P, Treister N, Baggott T, Gibson P, Tissing W, et al. Guideline for the prevention of oral and oropharyngeal mucositis in children receiving treatment for cancer or undergoing haematopoietic stem cell transplantation. BMJ Support Palliat Care. 2015 Mar 27.
[32] Robinson PD, Oberoi S, Tomlinson D, Duong N, Davis H, Cataudella D, et al. Management of fatigue in children and adolescents with cancer and in paediatric recipients of haemopoietic stem-cell transplants: a clinical practice guideline. Lancet Child Adolesc Heal. 2018 May 1;2(5):371–8.
[33] Lehrnbecher T, Robinson P, Fisher B, Alexander S, Ammann RA, Beauchemin M, et al. Guideline for the Management of Fever and Neutropenia in Children With Cancer and Hematopoietic Stem-Cell Transplantation Recipients: 2017 Update. J Clin Oncol. 2017 Jun 20;35(18):2082–94.
[34] International Pediatric Oncology Guidelines in Supportive Care Network (iPOG Network) [Internet]. [cited 2018 Aug 14]. Available from: http://www.sickkids.ca/ Research/iPOG/
[35] Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck- Y, Alonso-coello P, et al. GRADE : An Emerging Consensus on Rating Quality of Evidence a nd Strength of Recommendations. 2008;336(April):924–6.
[36] Dutch National Health Care Institute. Advies- en expertgroep Kwaliteitsstandaarden (AQUA) - Leidraad voor kwaliteitsstandaarden [Internet] . [cited October 1st 2017]
. Available from: https://www.zorginzicht.nl/kennisbank/Documents/Leidraad voor kwaliteitsstandaarden %28AQUA%29.PDF
[37] Guyatt GH, Oxman AD, Kunz R, Atkins D, Brozek J, Vist G, et al. GRADE guidelines: 2. Framing the question and deciding on important outcomes. J Clin Epidemiol. 2011;64(4):395–400.
29 GENERAL INTRODUCTION
[38] Alonso-Coello P, Schünemann HJ, Brignardello-Petersen R, Akl EA, Davoli M, Treweek S, et al. GRADE Evidence to Decision (EtD) frameworks: a systematic and transparent approach to making well informed healthcare choices. 1: Introduction. BMJ. 2016;353(11):i2016.
[39] Alonso-Coello P, Oxman AD, Moberg J, Brignardello-Petersen R, Akl EA, Davoli M, et al. GRADE Evidence to Decision (EtD) frameworks: a systematic and transparent approach to making well informed healthcare choices. 2: Clinical practice guidelines. BMJ. 2016;353(11):i2089.
[40] Schünemann H, Brozek J, Guyatt G, Oxman A. GRADE Handbook for grading quality of evidence and strength of recommendation. 2010.
[41] Brusamento S, Legido-Quigley H, Panteli D, Turk E, Knai C, Saliba V, et al. Assessing the effectiveness of strategies to implement clinical guidelines for the management of chronic diseases at primary care level in EU Member States: a systematic review. Health Policy. 2012 Oct;107(2–3):168–83.
[42] Carlbom DJ, Rubenfeld GD. Barriers to implementing protocol-based sepsis resuscitation in the emergency department—Results of a national survey. Crit Care Med. 2007;35(11):2525–32.
[43] Grol R, Grimshaw J. From best evidence to best practice: effective implementation of change in patients’ care. Lancet. 2003;362:1225–30.
[44] Prior M, Guerin M, Grimmer-Somers K. The effectiveness of clinical guideline implementation strategies--a synthesis of systematic review findings. J Eval Clin Pract. 2008 Oct;14(5):888–97.
[45] Lawrence M, Olesen F. Indicators of Quality in Health Care. Eur J Gen Pract. 2009;3(3):103–8.
[46] Webber C, Siemens DR, Brundage M, Groome PA. Quality of care indicators and their related outcomes: A population-based study in prostate cancer patients treated with radical prostatectomy. Can Urol Assoc J. 2014 Jul;8(7–8):E572-9. [47] Armes J, Wagland R, Finnegan-John J, Richardson A, Corner J, Griffiths P.
Development and Testing of the Patient-Reported Chemotherapy Indicators of Symptoms and Experience. Cancer Nurs. 2014;37(3):E52–60.
[48] Ouwens M, Hermens R, Hulscher M, Vonk-Okhuijsen S, Tjan-Heijnen V, Termeer R, et al. Development of indicators for patient-centred cancer care. Support Care Cancer. 2010 Jan;18(1):121–30.
[49] De Vos M, Graafmans W, Kooistra M, Meijboom B, Van Der Voort P, Westert G. Using quality indicators to improve hospital care: a review of the literature. Int J Qual Heal Care. 2009 Apr 1;21(2):119–29.
P
h D
P
+
1
preparation
for guideline
development
PART 1
CHAPTER 2
development of clinical practice
guidelines for supportive care in
childhood cancer – prioritization
of topics using a Delphi approach
Published as: Loeffen EAH Mulder RL Kremer LCM Michiels EMC Abbink FC Ball LM Segers H Mavinkurve-Groothuis AM Smit FJ Vonk IJ van de Wetering MD Tissing WJE Support Care Cancer. 2015 Jul;23(7):1987-95.
34 CHAPTER 2
2.1 ABSTRACT
INTRODUCTION
Currently, very few guidelines for supportive care for children with cancer exist. In the Netherlands, nationwide guidelines are over 10 years old and mostly based on expert opinion. Consequently, there is growing support and need for clinical practice guidelines (CPGs), which ought to be developed with a well-defined methodology and include a systematic search of literature, evidence summaries and a transparent description of the decision process for the final recommendations. Development of CPGs is time consuming, therefore it is important to prioritize topics for which there is the greatest clinical demand.
OBJECTIVES
To prioritize childhood cancer supportive care topics for development of CPGs.
METHODS
A Delphi survey consisting of two rounds was conducted to prioritize relevant childhood cancer supportive care topics for the development of CPGs. A group of experts comprising 15 pediatric oncologists, 15 pediatric oncology nurses and 15 general pediatricians involved in care for childhood cancer patients were invited to participate. All relevant supportive care topics in childhood cancer were rated.
RESULTS
In both rounds, 36 panelists (82%) responded. Agreement between panelists was very good, with an intraclass correlation coefficient of 0.918 (95% CI = 0.849-0.966, p<0.001) in round 2. The 10 topics with the highest score in the final round were infection, sepsis, febrile neutropenia, pain, nausea/vomiting, restrictions in daily life and activities, palliative care, procedural sedation, terminal care and oral mucositis.
CONCLUSION
We successfully used a Delphi survey to prioritize childhood cancer supportive care topics for the development of CPGs. This is a first step towards uniform and evidence-based Dutch guidelines in supportive care in childhood cancer. Even though performed nationally, we believe that this study can also be regarded as an example starting point for international development of CPGs in the field of supportive care in cancer, or any other field for that matter.
35 PRIORITIZING SUPPORTIVE CARE TOPICS
2.2 INTRODUCTION
The current high survival rates for children with cancer of 75%-80% are the results of the introduction of intensive treatment protocols consisting of surgery, radiotherapy and chemotherapy.[1] However, some of these treatment strategies cause severe adverse effects that can lead to treatment-related death. A study among children with acute lymphoblastic leukemia found treatment-related death to be ranked as the second leading cause of death, after cancer itself.[2] In addition, adverse effects might lead to delay of treatment or dose reduction of chemotherapy, which causes suboptimal treatment and might thus also lower survival chances. Moreover, as shown in children with acute myeloblastic leukemia, these adverse effects might lead to vast morbidity and a decreased quality of life.[3] Supportive care focuses on the prevention and management of these adverse effects of cancer and its treatment. Thus, optimal supportive care is needed to reduce and/or prevent morbidity and mortality and to improve quality of life for children with cancer and their families.[4,5]
To ensure that childhood cancer patients receive optimal care, CPGs are essential.[6,7] CPGs are defined as “statements that include recommendations intended to optimize patient care that are informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options.”.[8] CPGs aim to bridge the gap between research and clinical practice and are regarded as powerful tools to improve the quality of care.[9-11] In addition, CPGs can contribute to a reduced variability in daily practice and costs.[6,7,12]
In the Netherlands national guidelines for supportive care are over 10 years old and are mostly based on expert opinion.[13] Every clinic and sometimes even every individual care provider can make their own choices regarding optimal supportive care interventions. This can result in conflicting recommendations for healthcare providers, children with cancer and their parents and thus in suboptimal care. In addition, clinical practice variation is associated with increased health care costs.[14]
Internationally, evidence based guidelines regarding supportive care in childhood cancer are sparse as well. The few that exist are aimed at specific areas of the major supportive care topics, e.g. nausea and vomiting, and febrile neutropenia.[15-18] However, from personal experience and contacts with colleagues abroad we know that the will to develop CPGs regarding supportive care in childhood cancer not merely exists in the Dutch Childhood Oncology Group (NL), but also in multiple international guideline groups, among others the Children’s Oncology Group (US), the C17 Council (CA), the
36
CHAPTER 2
Pediatric Oncology Group of Ontario (CA), the Children’s Cancer and Leukaemia Group (GB) and the Swiss Paediatric Oncology Group (CH).
For guidelines to be of optimal use, it is of utmost importance to develop them via a vast methodology, base them on summaries of the best available evidence, clearly distinct between weak and strong recommendations and include appropriate recognition for shared decision making, i.e. acknowledge patient values and preferences.[19]
Therefore, we have initiated a project for the development of CPGs for supportive care in childhood cancer. As the development of CPGs is time consuming and supportive care is a very extensive field, a first step was to identify and prioritize supportive care topics for which there is the greatest clinical demand. To achieve this, we conducted a Delphi survey among professionals in the field of childhood cancer.
2.3 METHODS
DELPHI APPROACH
The Delphi approach is a well-known method for reaching consensus in a group and relies on anonymity, iteration, controlled feedback and statistical group response.[14] In multiple rounds incorporating feedback of the groups’ response, participants compile a consensus list.
SELECTION OF CORE RESEARCH TEAM
Before commencing this study, a core research group was formed to guide the process of the Delphi survey. This group comprised seven members: two pediatric oncologists, one pediatrician trained in epidemiology, one pediatrician involved in care for childhood cancer patients in a shared care center, one pediatric oncology nurse, one postdoctoral scholar and one PhD student (MD). The core research group formed the initial list of relevant topics; each member was asked to individually form a list of all supportive care topics he or she could think of, subsequently these were all incorporated into the final list. Also, the core research group decided on the number of rounds of the survey and discussed the outcome of each round.
SELECTION OF PANELISTS
For this Delphi survey, we invited various professionals from all over The Netherlands; 15 pediatric oncologists and 15 pediatric oncology nurses from all seven Dutch childhood oncology centers. Furthermore, 15 general pediatricians involved in care for childhood cancer patients from various Dutch shared care centers were invited.
37 PRIORITIZING SUPPORTIVE CARE TOPICS
DELPHI SURVEY
The list of relevant topics, as formed by the core research team, was incorporated in a Delphi survey to determine the order of development of CPGs for all topics in supportive care in childhood cancer. This survey was sent to all panelists by e-mail, with the option to send it back digitally or per mail. In both rounds, two weeks after the initial distribution e-mail reminders were sent to all non-responders, two weeks later we attempted to contact all non-responders personally, by phone or by e-mail.
Figure 2.1. Outline of the Delphi approach.
In Figure 2.1 the Delphi approach is shown. In the first Delphi round, panelists were asked to rate prevalence, severity and adequate treatment options for each supportive care topic on a 5-point Likert scale ranging from one (low ranking) to five (high ranking). The panelists also had the option to provide written comments and to suggest missing topics (see Supplemental material 2/S1). For the second Delphi round, only topics were selected that had mean scores of >2.5 on all three questions in round one. This was chosen because we wanted to start supportive care guideline development with topics that are clinically most relevant, which we regarded as occurring often, being relatively severe and having adequate treatment options. Of these, the 20 highest scoring topics were selected and presented in descending order in the second Delphi round (see
38 CHAPTER 2
Supplemental material 2/S1). This was limited to the 20 highest scoring items as we wanted our questionnaires to be focused on the most important topics to start CPG development with, and minimize the risk of causing fatigue (and thus perhaps non-response) in our panelists. Also, all written additions were included. In this second round the panelists were asked to rate how important it is that a CPG will be developed for the specific topic, on a 5-point Likert scale ranging from one (unimportant) to five (very important). Before commencing our survey, the core research group decided to initially limit the number of rounds to two, as this Delphi survey had a prioritizing rather than a selective character. We decided to consider more rounds when there was no agreement between panellists. Each questionnaire took less than 15 minutes to complete.
ANALYSIS
We explored agreement between panelists. We determined there was agreement when more than 66% of all panelists rated a four or five on the 5-point Likert of a specific topic. In addition, the level of agreement between panelists was estimated with the intraclass correlation coefficient (ICC), set to a two-way mixed model with absolute agreement. [20] As our study involved several panelists, we present the average measure ICC, with 95% confidence interval (CI) and p-value. Cut-off values for ICC are arbitrary, but similar to kappa statistics agreement was categorized as poor (ICC <0.2), fair (ICC 0.3-0.4), moderate (ICC 0.5-0.6), good (ICC 0.7-0.8) or very good (ICC >0.8).[21]
Statistical analyses were performed using IBM SPSS Statistics version 22.0 (International Business Machines Corporation, NY, USA).
2.4 RESULTS
Between July 2013 and December 2013, panelists completed a series of two questionnaires. A total of 45 experts were invited to participate of whom 44 responded and were willing to do so. Round one of the Delphi survey consisted of 41 topics and comprised the full initial list of supportive care topics, as formed by the core research group (Table 2.1). The response rate in round one was 82% (36/44 experts). Five experts passively dropped out of the panel after round one, i.e. did not respond to several reminders for round two. In the second round, five experts that did not return round one did participate in this round. Thus, the response rate in round two was also 82% (36/44 experts). In all, the panel returning round two consisted of 14 pediatric oncologists, 11
39 PRIORITIZING SUPPORTIVE CARE TOPICS
pediatric oncology nurses and 11 general pediatricians involved in care for childhood cancer patients.
Table 2.1. Results of the Delphi survey round one, sorted by descending overall mean score on all three items.
Mean score†
Supportive care topic Prevalence Severity Adequate
treatment options Overall mean score
Anemia‡ 4.50 2.89 4.72 4.04 Infection‡ 4.28 3.78 4.02 4.03 Nausea / vomiting‡ 4.69 3.25 4.02 3.99 Thrombocytopenia‡ 4.56 3.03 4.33 3.97 Pain‡ 4.06 3.72 4.06 3.94 Febrile neutropenia‡ 4.06 3.44 3.83 3.78 Constipation‡ 4.00 2.92 4.42 3.78 Sepsis‡ 2.78 4.67 3.72 3.72 Malnutrition‡ 3.81 3.22 4.00 3.68 Leukopenia‡ 4.67 3.69 2.42 3.59 Psychosocial issues‡ 4.06 3.60 3.00 3.55
Tumor lysis syndrome 2.41 3.79 4.15 3.45
Palliative care‡ 2.81 4.03 3.52 3.45
Hypertension‡ 3.00 3.03 4.22 3.42
Mucositis (oral) ‡ 3.69 3.47 3.00 3.39
Graft versus host disease after HSCT*‡ 3.00 4.15 3.00 3.38
Virus reactivation after HSCT*‡ 3.04 4.04 3.04 3.37
Varicella virus‡ 2.78 3.47 3.77 3.34
Terminal care‡ 2.64 4.09 3.27 3.33
Mucositis (gastrointestinal) ‡ 3.25 3.69 2.92 3.29
Endocrine complications‡ 2.68 3.33 3.68 3.23
PAC / VAP*; occlusion 2.83 3.23 3.60 3.22
Thrombosis 2.31 3.80 3.51 3.21 Menstruation 3.06 2.56 3.94 3.19 Sub-/infertility 3.19 4.36 2.03 3.19 Fatigue 4.11 3.17 2.28 3.19 Sinus thrombosis 1.94 4.17 3.43 3.18 Diabetes mellitus 2.17 3.34 3.97 3.16 Pancreatitis 2.00 4.25 3.06 3.10 Neutropenic colitis 2.00 4.06 3.09 3.05
Nephrological complications; tubular 2.64 3.21 3.24 3.03
Hyperviscosity syndrome 1.43 3.47 4.03 2.98
Superior vena cava syndrome 1.46 3.94 3.50 2.97
40 CHAPTER 2
Table 2.1. Continued
Mean score†
Supportive care topic Prevalence Severity Adequate
treatment options Overall mean score Posterior reversible encephalopathy
syndrome
1.86 3.97 2.97 2.93
Skin defect after radiotherapy 2.45 3.30 2.88 2.88
Cardiomyopathy 1.69 4.26 2.57 2.84 Nephrological complications; glomerular 2.12 3.33 2.97 2.81 Veno-occlusive disease 1.47 4.21 2.63 2.77 Disruption of taste 3.94 2.56 1.78 2.76 Extravasation of chemotherapy 1.36 3.91 2.77 2.68 Fever 2.58 2.83 2.52 2.65
† On a 5-point Likert scale ranging from 1 (low ranking) to 5 (high ranking). ‡ These items were included in round 2.
* PAC = port-a-cath, VAP = venous access port, HSCT = hematopoietic stem cell transplantation.
Table 2.2. Results of the Delphi survey round two, sorted by descending mean scores. Importance to develop a CPG*
Supportive care topic Mean score† Median score† Range†
Infection 4.61 5 3 – 5
Sepsis 4.33 5 2 – 5
Febrile neutropenia 4.28 5 2 – 5
Pain 4.17 4 2 – 5
Nausea / vomiting 4.14 4 2 – 5
Restrictions in daily life and activities 4.06 4 2 – 5
Palliative care 3.91 4 1 – 5 Procedural sedation 3.86 4 2 – 5 Terminal care 3.83 4 1 – 5 Mucositis (oral) 3.75 4 2 – 5 Varicella virus 3.64 4 2 – 5 Mucositis (gastrointestinal) 3.64 4 2 – 5 Malnutrition 3.61 4 2 – 5 Thrombocytopenia 3.50 4 1 – 5 Peripheral neuropathy 3.39 3 2 – 5 Anemia 3.33 3 1 – 5
Graft versus host disease after HSCT* 3.31 3 1 – 5
Leukopenia 3.29 3 2 – 5
Virus reactivation after HSCT* 3.23 3 1 – 5
41 PRIORITIZING SUPPORTIVE CARE TOPICS Table 2.2. Continued
Importance to develop a CPG*
Supportive care topic Mean score† Median score† Range†
Allergic reactions 3.19 3 1 – 5 Constipation 3.08 3 1 – 5 Hypertension 3.08 3 1 – 5 Endocrine Complications 3.03 3 1 – 5 Psychosocial issues 2.94 3 2 – 5 Hemorrhagic cystitis 2.89 3 1 – 4
Lung function disorder 2.86 3 1 – 5
Ototoxicity 2.72 3 1 – 4
Hypertriglyceridemia 2.69 3 1 – 5
Alopecia 2.36 2 1 – 5
† On a 5-point Likert scale ranging from 1 (low ranking) to 5 (high ranking).
* CPG = clinical practice guideline, HSCT = hematopoietic stem cell transplantation, ANFH = avascular necrosis of the femoral head.
Note: items highlighted in italic were written additions.
The mean scores for each supportive care topic of round one are presented in Table 2.1. In round one, 16 topics had mean scores below 2.5 on one of the three questions and were therefore eliminated (see Table 2.1). The 20 highest scoring topics were included in round two, as well as all written additions. There were 10 written additions: allergic reactions, alopecia, hemorrhagic cystitis, hypertriglyceridemia, lung function disorder, osteoporosis / avascular necrosis of the femoral head, ototoxicity, procedural sedation, peripheral neuropathy, and restrictions in daily life and activities.
After Delphi round two the panelists determined the following topics to be prioritized in upcoming development of CPGs: 1) infection, 2) sepsis, 3) febrile neutropenia, 4) pain, 5) nausea / vomiting, 6) restrictions in daily life and activities, 7) palliative care, 8) procedural sedation, 9) terminal care and 10) oral mucositis (Table 2.2).
Regarding the first nine topics of this top 10 there was agreement between panelists, i.e. more than 66% of all panelists rated a four or five on the 5-point Likert scale of that topic. Few panelists scored these items with a one or two, ranging from 0% (infection) to 14.3% (procedural sedation). With regard to the level of agreement, the intraclass correlation coefficient of the 20 topics presented in both round one and two was 0.904 (95% CI = 0.824-0.960, p<0.001) for round one and 0.918 (95% CI = 0.849-0.966, p<0.001) for round two. Thus, in both rounds agreement between panelists was very good, with a minor increase from round one to round two.
