Needs with Regard to Decision Support Systems for Treating Patients with Incurable Non-small Cell Lung Cancer

  • Dóra RévészEmail author
  • Ellen G. Engelhardt
  • Johannes J. Tamminga
  • Franz M. N. H. Schramel
  • Bregje D. Onwuteaka-Philipsen
  • Ewoudt M. W. van de Garde
  • Ewout W. Steyerberg
  • Henrica C.W. de Vet
  • Veerle M. H. Coupé
Open Access


Treatment decision-making for patients with incurable non-small cell lung cancer (NSCLC) is complex due to the rapidly increasing number of treatments and discovery of new biomarkers. Decision support systems (DSS) could assist thoracic oncologists (TO) weighing of the pros and cons of treatments in order to arrive at an evidence-based and personalized treatment advice. Our aim is to inventory (1) TO’s needs with regard to DSS in the treatment of incurable (stage IIIB/IV) NSCLC patients, and (2) preferences regarding the development of future tools in this field. We disseminated an online inventory questionnaire among all members of the Section of Oncology within the Society of Physicians in Chest Medicine and Tuberculosis. Telephone interviews were conducted to better contextualize the findings from the questionnaire. In total, 58 TO completed the questionnaire and expressed a need for new DSS. They reported that it is important for tools to include genetic and immune markers, to be sufficiently validated, regularly updated, and time-efficient. Also, future DSS should incorporate multiple treatment options, integrate estimates of toxicity, quality of life and cost-effectiveness of treatments, enhance communication between caregivers and patients, and use IT solutions for a clear interface and continuous updating of tools. With this inventory among Dutch TO, we summarized the need for new DSS to aid treatment decision-making for patients with incurable NSCLC. To meet the expressed needs, substantial additional efforts will be required by DSS developers, above already existing tools.


Decision support systems Non-small-cell lung cancer Surveys and questionnaires 


Lung cancer, of which roughly 80–85% is non-small cell lung cancer (NSCLC) [1], is one of the leading causes of cancer mortality worldwide [2]. Especially, incurable NSCLC patients (stages IIIB and IV) have a very short life expectancy with 1-year survival probabilities ranging between 22 and 47% [3]. Palliative treatment in this setting aims to preserve or improve quality of life, lengthen life, or decrease disease burden. Palliation can consist of symptom relief (best supportive care), or can target tumor tissue itself with, for instance, systemic chemotherapy (CT), radiotherapy (RT), targeted therapies (e.g., epidermal growth factor receptor tyrosine kinase inhibitors, EGFR-TKI), or immunotherapies.

Individually tailored palliative cancer treatment is essential to ensure that patients receive the treatment that optimally matches their values and preferences, avoiding under- or overtreatment, and optimally utilizing all available healthcare resources. Decision-making in the palliative phase is complex due to the large heterogeneity in this population with regard to histological tumor type, biomarker profile, treatment response, and the marginal expected treatment benefit and the potentially severe toxicity of systemic therapy. Particularly with the rapid developments and breakthroughs in the fields of genomic technologies and molecular therapies for lung cancer, treatment decisions will become more multi-factorial and personalized, and it will become increasingly important to support thoracic oncologists (TO) in staying up-to-date.

Treatment guidelines are quite extensive and general, and do not allow for individual tailoring of treatment. Decision support systems (DSS), as add-ons to existing guidelines, can assist TO in formulating an evidence-based treatment advice by comparing various treatment options for an individual patient. Such tools ideally weigh the pros (e.g., symptom relief, life lengthening) and cons of treatments (e.g., side effects, loss of quality of life), as well as the cost-effectiveness. DSS can consist of (multiple) prediction models, nomograms, and decision trees, which are often based on patient and tumor characteristics. Multiple DSS have been developed for incurable NSCLC patients [4], such as the prognostic score developed by Florescu et al. [5] and the nomograms developed by Hoang et al. [6]. These tools are developed to decide whether certain treatments bring enough gain in overall or progression-free survival to be worthwhile for patients. Research has shown that oncologists tend to overestimate life expectancy of patients [7, 8], and that they could benefit from using statistical tools providing survival estimates [7, 9, 10, 11].

Although there are a large number of DSS available to aid palliative treatment decisions for incurable NSCLC patients [4], it remains unknown whether the needs of TO are met by the existing tools. Our aim was to shed light on the unmet needs and preferences of TO for future tools in this rapidly developing field.


Online Inventory Questionnaire: Development and Dissemination

We developed an online inventory questionnaire (using the Surveymonkey platform). The inventory first covered background information of the respondents, the available DSS and clinical parameters that they use in decision-making, and whether they have access to a palliative care team in the hospital. Next, we inventoried what the decisive reasons are for not using a DSS, whether they express a need for future tools, and what they would require in terms of future DSS for palliative care decision-making for patients with incurable NSCLC. The online inventory questionnaire was disseminated among the members (i.e., 157 TO) of the Section of Oncology (SON) within the Society of Physicians in Chest Medicine and Tuberculosis (NVALT). The research protocol was approved by the ethical committee of the VU University Medical Center and all respondents provided written informed consent.

Expert Interviews

Respondents who indicated in the inventory that they were willing to provide additional information about their DSS needs were contacted. Semi-structured phone interviews were then used to explore TO’s needs and preferences concerning (future) DSS in more detail. The aim of the interviews was to put the findings from the questionnaire in context.

Data Extraction and Analyses

Descriptive analyses were performed on the inventory data. The interviewer (DR) extracted key points from the telephone interviews, and incorporated them into the data from the inventory questionnaire.


Respondent Characteristics and Use of DSS or Clinical Parameters

In total, 58 TO filled out the online inventory, of which 37 completed the entire questionnaire (Table 1). The majority of our respondents reported that they treat > 50 incurable NSCLC patients per year, and had > 10 years of experience treating incurable NSCLC patients. Most respondents worked in academic medical centers, and had access to a palliative care team at their hospital. Of the 58 responding TO, 15 were willing to be contacted for in-depth interviews, of which four TO were reached for interviews. In total, 31 TO reported that they make use of an existing tool, although they mostly used clinical guidelines (Table 1). Only few respondents used other DSS, such as the Florescu et al. prognostic score or the recursive partitioning analysis score. Furthermore, they take a wide range of clinical parameters into account during decision-making in their daily practice, such as patients’ performance status, ability to cope, preferences, and the impact on quality of life (Fig. 1).
Table 1

General characteristics of thoracic oncologists that completed the online survey (N = 58)

Work setting

N = 58*

 Academic medical centers and teaching hospitals

25 (43%)

 General hospital

21 (36%)

 University medical center

8 (14%)

 Specialized oncological center

3 (5%)

 Top clinical care hospitals

1 (2%)

Experience with incurable non-small cell lung cancer patients

N = 58



 ≤ 2 years

3 (5.2%)

 3–5 years

7 (12.1%)

 6–10 years

14 (24.1%)

 > 10 years

34 (58.6%)

Number of incurable non-small cell lung cancer patients per year

N = 58

 ≤ 5 patients


 6–10 patients

1 (1.7%)

 11–20 patients

1 (1.7%)

 21–50 patients

15 (25.9%)

 > 50 patients

41 (70.7%)

Access to palliative care team in hospital?

N = 58


57 (98.3%)

Use of existing DSS?

N = 33*


31 (93.9%)

 Dutch national NSCLC guidelines

23 (69.7%)

 Dutch national palliative care guidelines

16 (48.5%)

 Florescu et al. prognostic score

2 (6.1%)

 Recursive partitioning analysis score

2 (6.1%)

 National Comprehensive Cancer Network guidelines

2 (6.1%)

 Hoang et al. nomograms for survival


 Supportive and Palliative Care Indicators Tool (SPICT)


Need of new DSS?

N = 51*


39 (76.5%)

 Yes, for shared decision-making

20 (39.2%)

 Yes, for treatment decision-making

4 (7.8%)

 Yes, for decision-making and shared decision-making

15 (29.4%)

*Numbers can vary due to different amounts of responders, and they do not always add up to 100% as multiple answers were allowed

Fig. 1

Clinical factors for decision-making in the treatment of incurable NSCLC patients (N = 56)

Needs and Preferences for (Future) DSS

With the online inventory and the in-depth interviews, we aimed to shed light on the needs and preferences of TO who treat incurable NSCLC patients. Figure 2 summarizes the factors that have been reported to determine whether they use an existing DSS or not. Firstly, DSS should be sufficiently validated and evidence-based, up-to-date (requiring regular updates), and should not be time-consuming. Furthermore, DSS should be easy to use with a clear interface, easy to explain to patients, and freely available, preferably as an online or mobile application.
Fig. 2

Factors reported to be important for the choice of using certain DSS by TO (N = 51)

In addition to the existing tools, 39 TO reported that they would welcome new DSS for their own decision-making and/or for shared decision-making (Table 1). When we asked TO about the specific preferences for future DSS, there was a broad variety of unmet needs reported. These are summarized in Fig. 3, and classified into the aspects “treatments,” “integration,” “communication,” and “using IT solutions.” Firstly, needs were expressed for specific treatment decisions, such as the continuation or discontinuation of specific therapies. Also, future DSS should integrate different treatment options, tailored to the individual patient in a clear overview. These tools should then not only focus on benefits (such as survival), but also the harms of treatment (e.g., toxicity), and on other societally relevant outcome (such as cost-effectiveness). Furthermore, DSS should function as communication enhancers, both between professionals from different disciplines and with patients. Lastly, information technology (IT) can aid by (1) creating clear interfaces, and (2) facilitating information provision to patients about the expected treatment benefits and the side effects of the various treatment options. Also, IT can support future DSS by offering sophisticated algorithms for continuous updating of DSS, and by, for example, incorporating DSS into the electronic patient files. Especially in the telephone interviews, TO mentioned that DSS should not lead to an automated decision-making process, but merely present the advantages and disadvantages of the available treatment options. In other words, oncologists ought to maintain a central role in clinical decision-making.
Fig. 3

Needs and preferences of TO grouped into the categories Treatments, Integration, Communication, and Information Technology (IT) that can be used for development of future DSS


Online inventory questionnaires were filled out by 58 Dutch TO about their needs and preferences regarding DSS for the treatment of incurable NSCLC patients. The majority of the respondents reported that there is a need for newly developed, up-to-date and validated DSS. Based on the results of the current inventory, recommendations can be made for future DSS development. For optimal use in clinical practice, such tools should integrate all relevant treatment options, weigh the benefits and harms of treatments, facilitate communication between caregivers and with patients, and be kept up-to-date.

To support TO’s decision-making, DSS are a useful and time-efficient complement to clinical guidelines for treatment decision-making. Guidelines could even incorporate DSS that are ready for use in clinical practice. Currently, a large number of DSS have been developed for this population, as shown in our previously published systematic review [4]. Even though the TO in our sample used some of these existing DSS, they reported a need for new tools in their daily practice for multiple goals. Firstly, they missed DSS that show for each individual patient which (combination of) treatments are appropriate, and what the estimated survival gain and toxicities are for each treatment option. Secondly, data on available treatment options, and knowledge from clinical trials, real-life data, and guidelines should be integrated into a DSS. Thirdly, the communication among the different caregivers and with patients could be improved. Lastly, IT solutions were deemed essential, as the challenge will be to design methods for continuous updating of DSS. Moreover, creating clear interfaces for future DSS would make them more suitable to assist in shared decision-making.

Various separate factors, such as performance status, ability to cope, preferences, and the impact on quality of life were reported to be important in clinical decision-making. These are often already incorporated in DSS, but are measured relatively subjectively. Future DSS will need more objective markers, such as genetic and serum markers, in order to reduce unwanted treatment variety and correct use of expensive medication. In the past decades, great progress has been made in the understanding of underlying biological processes of cancer cells. This has led to the discovery of several onco-pathways, which are controlled by oncogenes and tumor suppressor genes. Biological markers are becoming highly influential prognostic and predictive factors for clinical decision-making, as they determine the effectiveness of targeted and immune therapies. In adenocarcinoma and squamous cell carcinoma, various important onco-pathways (e.g., mitogen-activated protein kinase and phosphoinositide 3-kinase pathways) [12, 13], oncogenes (e.g., KRAS, EGFR, BRAF), tumor suppressor genes (e.g., TP53, STK11, CDKN2A), and gene rearrangements (e.g., ALK, ROS, and RET) have been found to be important [14, 15, 16]. Furthermore, angiogenesis and the vascular endothelial growth factor affect the processes of primary tumor growth, proliferation, and metastasis [17]. In addition, the immune checkpoints programmed cell death protein-1 and programmed cell death ligand-1 pathways have recently emerged as important targets for immunotherapy [18, 19]. Currently, existing DSS are not up-to-date with regard to such biological markers. Therefore, future tools should incorporate genetic tumor profiles and serum markers to get optimal selection of patients for the myriad of treatment options available. Currently, various national and international guidelines mainly include information about certain genetic mutations of, for example, EGFR and KRAS, or rearrangements of, for example, ALK and ROS1 [20, 21, 22, 23], but there is not a clear decision tool that is tailored to individual patients.

To overcome the reported limitations and to create an adequate tool for future clinical practice, various steps have to be undertaken. It is obvious that oncologist cannot keep up with the current literature to choose the optimal evidence-based treatments, based on a balance between survival gain, side effects, quality of life, patient preferences, and even cost-effectiveness. In addition, the wide array of factors that influence clinical decision-making renders it too complex for the human cognitive capacity. In general, guidelines based on systematic reviews that summarize the evidence of pros and cons of the various treatment options are very useful. Regular updates of the existing guidelines are necessary, but apart from these, DSS can give guidance to decide on the optimal treatment of individual patients. Using the infrastructure of electronic health records for clinical DSS could support personalized medicine practices by providing the tailored information at the right time. Various researchers have discussed the advantages of incorporating DSS in electronic health records in combination with smart algorithms, such that continuous updating and rapid adaptation of DSS are ensured [11, 24, 25, 26, 27]. This precision oncology should be a multidisciplinary effort incorporating information from various experts and stakeholders [24, 25].

There are some ethical issues that were raised in conversations with TO, which should be kept in mind as well. A DSS can support the oncologist in giving treatment advice based on the best available evidence. If the DSS is available online and easy to understand, the given information about pros and cons of treatments can be visualized for patients. Nevertheless, this decision-making process should not be automated, even when good DSS become available; oncologists remain essential in the decision-making process. Furthermore, although the oncologist is offering the best treatment options based on available evidence and clinical knowledge, ultimately, the patient decides. However, patients might not feel capable of choosing between the available treatment options.

The strengths of this study were the fact that we had access to the entire Section of Oncology (SON) of the group of Physicians in Chest Medicine and Tuberculosis organization (largest TO organization in the Netherlands), and these TO were distributed over various hospital settings. Considering the fact that the vast majority of our sample used existing DSS, and reported a need for new tools, it is likely that TO with unmet needs have been more inclined to participate than the oncologists who have less knowledge about or need for DSS. Nevertheless, the current group of responders provided valuable insight in TO’s needs and preferences for new DSS.


In conclusion, TO in the palliative settings reported a need for newly developed DSS when it comes to clinical decision-making in the palliative phase of NSCLC. DSS can be an aid in clinical practice by summarizing and visualizing the current fast-moving area of biomarker-based palliative lung cancer treatments. Future DSS should include novel important prognostic and predictive factors, such as genetic and immune markers, and provide estimates for both survival rates, toxicity, and cost-effectiveness of the available treatments. Future DSS developers should find methods to ensure the continuous updating of DSS and their implementation into daily clinical practice.



Firstly, we would like to thank Dr. Annemarie Becker-Commissaris for her valuable input in the design of the online inventory questionnaires for incurable NSCLC patients. Furthermore, we would like to thank all participating thoracic oncologists who took the time to fill out the online questionnaire, and for taking the time to participate in the telephone interviews. This research was commissioned by The National Health Care Institute (ZIN: Zorginstituut Nederland).


  1. 1.
    American Cancer Society [updated 01-07-2016; cited 2016 23-02-2016]. Available from: Accessed 1 July 2016
  2. 2.
    World Health Organization Cancer Fact sheet [updated 01-02-2015; cited 2016 23-02-2016]. Available from: Accessed 23 Feb 2016
  3. 3.
    Incidentie-, prevalentie- en sterftecijfers: Nederlandse Kankerregistratie beheerd door IKNL; 2016 [updated 04-02-2016; cited 2016 23-02-2016]. Available from: Accessed 4 Feb 2016
  4. 4.
    Revesz D, Engelhardt EG, Tamminga JJ, Schramel F, Onwuteaka-Philipsen BD, van de Garde EMW et al (2017) Decision support systems for incurable non-small cell lung cancer: a systematic review. BMC Med Inform Decis Mak 17(1):144CrossRefGoogle Scholar
  5. 5.
    Florescu M, Hasan B, Seymour L, Ding K, Shepherd FA (2008) A clinical prognostic index for patients treated with erlotinib in National Cancer Institute of Canada Clinical Trials Group study BR.21. J Thorac Oncol 3(6):590–598CrossRefGoogle Scholar
  6. 6.
    Hoang T, Dahlberg SE, Sandler AB, Brahmer JR, Schiller JH, Johnson DH (2012) Prognostic models to predict survival in non-small-cell lung cancer patients treated with first-line paclitaxel and carboplatin with or without bevacizumab. J Thorac Oncol 7(9):1361–1368CrossRefGoogle Scholar
  7. 7.
    Tseng YD, Krishnan MS, Sullivan AJ, Jones JA, Chow E, Balboni TA (2013) How radiation oncologists evaluate and incorporate life expectancy estimates into the treatment of palliative cancer patients: a survey-based study. Int J Radiat Oncol Biol Phys 87(3):471–478CrossRefGoogle Scholar
  8. 8.
    Chiu N, Chiu L, Lutz S, Zhang N, Lechner B, Pulenzas N, Popovic M, DeAngelis C, Milakovic M, Chow R, Chow E (2015) Incorporation of life expectancy estimates in the treatment of palliative care patients receiving radiotherapy: treatment approaches in light of incomplete prognostic models. Ann Palliat Med 4(3):162–168PubMedGoogle Scholar
  9. 9.
    Hallen SA, Hootsmans NA, Blaisdell L, Gutheil CM, Han PK (2015) Physicians’ perceptions of the value of prognostic models: the benefits and risks of prognostic confidence. Health Expect 18(6):2266–2277CrossRefGoogle Scholar
  10. 10.
    Engelhardt EG, Pieterse AH, van Duijn-Bakker N, Kroep JR, de Haes HC, Smets EM et al (2015) Breast cancer specialists’ views on and use of risk prediction models in clinical practice: a mixed methods approach. Acta Oncol 54(3):361–367CrossRefGoogle Scholar
  11. 11.
    Langton JM, Blanch B, Pesa N, Park JM, Pearson SA (2013) How do medical doctors use a web-based oncology protocol system? A comparison of Australian doctors at different levels of medical training using logfile analysis and an online survey. BMC Med Inform Decis Mak 13:82CrossRefGoogle Scholar
  12. 12.
    Bansal P, Osman D, Gan GN, Simon GR, Boumber Y (2016) Recent advances in targetable therapeutics in metastatic non-squamous NSCLC. Front Oncol 6:112PubMedPubMedCentralGoogle Scholar
  13. 13.
    Chen JY, Cheng YN, Han L, Wei F, Yu WW, Zhang XW, Cao S, Yu JP (2015) Predictive value of K-ras and PIK3CA in non-small cell lung cancer patients treated with EGFR-TKIs: a systemic review and meta-analysis. Cancer Biol Med 12(2):126–139PubMedPubMedCentralGoogle Scholar
  14. 14.
    Shea M, Costa DB, Rangachari D (2016) Management of advanced non-small cell lung cancers with known mutations or rearrangements: latest evidence and treatment approaches. Ther Adv Respir Dis 10(2):113–129CrossRefGoogle Scholar
  15. 15.
    Morgensztern D, Campo MJ, Dahlberg SE, Doebele RC, Garon E, Gerber DE, Goldberg SB, Hammerman PS, Heist RS, Hensing T, Horn L, Ramalingam SS, Rudin CM, Salgia R, Sequist LV, Shaw AT, Simon GR, Somaiah N, Spigel DR, Wrangle J, Johnson D, Herbst RS, Bunn P, Govindan R (2015) Molecularly targeted therapies in non-small-cell lung cancer annual update 2014. J Thorac Oncol 10(1 Suppl 1):S1–S63Google Scholar
  16. 16.
    Gandara DR, Hammerman PS, Sos ML, Lara PN Jr, Hirsch FR (2015) Squamous cell lung cancer: from tumor genomics to cancer therapeutics. Clin Cancer Res 21(10):2236–2243CrossRefGoogle Scholar
  17. 17.
    Hall RD, Le TM, Haggstrom DE, Gentzler RD (2015) Angiogenesis inhibition as a therapeutic strategy in non-small cell lung cancer (NSCLC). Transl Lung Cancer Res 4(5):515–523PubMedPubMedCentralGoogle Scholar
  18. 18.
    Herzberg B, Campo MJ, Gainor JF (2016) Immune checkpoint inhibitors in non-small cell lung cancer. OncologistGoogle Scholar
  19. 19.
    Tan WL, Jain A, Takano A, Newell EW, Iyer NG, Lim WT, Tan EH, Zhai W, Hillmer AM, Tam WL, Tan DSW (2016) Novel therapeutic targets on the horizon for lung cancer. Lancet Oncol 17(8):e347–e362CrossRefGoogle Scholar
  20. 20.
    Nederlandse Landelijke richtlijn Niet-kleincellig longcarcinoom, Versie: 2.3: Landelijke Werkgroep Longtumoren; 2015 [updated 19-12-2015; cited 2016 23-02-2016]. Available from: Accessed 23 Feb 2016
  21. 21.
    Schnipper LE, Davidson NE, Wollins DS, Tyne C, Blayney DW, Blum D, Dicker AP, Ganz PA, Hoverman JR, Langdon R, Lyman GH, Meropol NJ, Mulvey T, Newcomer L, Peppercorn J, Polite B, Raghavan D, Rossi G, Saltz L, Schrag D, Smith TJ, Yu PP, Hudis CA, Schilsky RL, American Society of Clinical Oncology (2015) American Society of Clinical Oncology statement: a conceptual framework to assess the value of cancer treatment options. J Clin Oncol 33(23):2563–2577CrossRefGoogle Scholar
  22. 22.
    Reck M, Popat S, Reinmuth N, De Ruysscher D, Kerr KM, Peters S (2014) Metastatic non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 25(Suppl 3):iii27–iii39CrossRefGoogle Scholar
  23. 23.
    Ettinger DS, Wood DE, Akerley W, Bazhenova LA, Borghaei H, Camidge DR, Cheney RT, Chirieac LR, D’Amico TA, Dilling TJ, Dobelbower MC, Govindan R, Hennon M, Horn L, Jahan TM, Komaki R, Lackner RP, Lanuti M, Lilenbaum R, Lin J, Loo BW Jr, Martins R, Otterson GA, Patel JD, Pisters KM, Reckamp K, Riely GJ, Schild SE, Shapiro TA, Sharma N, Stevenson J, Swanson SJ, Tauer K, Yang SC, Gregory K, Hughes M (2016) NCCN guidelines insights: non-small cell lung cancer, version 4.2016. J Natl Compr Cancer Netw 14(3):255–264CrossRefGoogle Scholar
  24. 24.
    Downing GJ, Boyle SN, Brinner KM, Osheroff JA (2009) Information management to enable personalized medicine: stakeholder roles in building clinical decision support. BMC Med Inform Decis Mak 9:44CrossRefGoogle Scholar
  25. 25.
    Johnson A, Zeng J, Bailey AM, Holla V, Litzenburger B, Lara-Guerra H, Mills GB, Mendelsohn J, Shaw KR, Meric-Bernstam F (2015) The right drugs at the right time for the right patient: the MD Anderson precision oncology decision support platform. Drug Discov Today 20(12):1433–1438CrossRefGoogle Scholar
  26. 26.
    Lambin P, Zindler J, Vanneste BG, De Voorde LV, Eekers D, Compter I, et al. Decision support systems for personalized and participative radiation oncology. Adv Drug Deliv Rev. 2016Google Scholar
  27. 27.
    Finlayson SG, Levy M, Reddy S, Rubin DL (2016) Toward rapid learning in cancer treatment selection: an analytical engine for practice-based clinical data. J Biomed Inform 60:104–113CrossRefGoogle Scholar

Copyright information

© The Author(s) 2019

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Dóra Révész
    • 1
    Email author
  • Ellen G. Engelhardt
    • 1
  • Johannes J. Tamminga
    • 1
  • Franz M. N. H. Schramel
    • 2
  • Bregje D. Onwuteaka-Philipsen
    • 3
  • Ewoudt M. W. van de Garde
    • 4
  • Ewout W. Steyerberg
    • 5
  • Henrica C.W. de Vet
    • 1
  • Veerle M. H. Coupé
    • 1
  1. 1.Department of Epidemiology and Biostatistics, Amsterdam Public Health Research InstituteVU University Medical CenterAmsterdamThe Netherlands
  2. 2.Department of Lung Diseases and TreatmentSt. Antonius HospitalNieuwegeinThe Netherlands
  3. 3.Department of Public and Occupational Health, Amsterdam Public Health Research InstituteVU University Medical CenterAmsterdamThe Netherlands
  4. 4.Department of Clinical PharmacySt. Antonius HospitalNieuwegeinThe Netherlands
  5. 5.Center for Medical Decision Sciences, Department of Public HealthRotterdamThe Netherlands

Personalised recommendations