Giving voice to patients to define treatment value

<p class="article-intro">The definition of the real value of a treatment for cancer patients should necessarily include the evaluation of their subjective experience, which can be measured with the adoption of patient-reported outcomes within clinical trials. Health-related quality of life (QoL) is universally considered a measure of clinical benefit, but its inclusion in clinical trials and in related publications is still suboptimal.</p> <p class="article-content"><div id="keypoints"> <h2>Keypoints</h2> <ul> <li>The assessment of QoL is crucial for a complete evaluation of the value of anticancer treatments.</li> <li>QoL is not assessed (or not reported) in a relevant proportion of phase III trials in cancer patients.</li> <li>In advanced/metastatic settings, due to symptoms and limited life expectancy, attention to QoL should be particularly high.</li> </ul> </div> <p>Many tools and questionnaires have been produced and validated to properly measure QoL in patients with lung cancer, as well as in other solid tumours. The inclusion of these measures in clinical trials has several important advantages. First of all, it allows providing a patient-focused assessment of the burden and the impact of disease. Moreover, it allows completely assessing the clinical benefit of a therapy by complementing the information provided by commonly used efficacy endpoints and the traditional description of safety based on physicians&rsquo; assessment. Furthermore, the availability of QoL data could allow a more accurate patient-physician communication about the treatments proposed, when they become accessible in clinical practice.<br /> We recently performed a systematic review in order to evaluate the adoption of QoL among endpoints in randomized phase III trials, published in recent years, testing anticancer drugs in solid tumors.¹ In the analysis, we included articles issued between 2012 and 2016 in 11 major journals where the majority of cancer phase III trials are usually published. In addition, we investigated deficiencies in the publication of QoL results (in terms of underreporting and delay of publication), considering both primary study publications and subsequent secondary publications including QoL results, when available.</p> <h2>QoL should be considered as trial endpoint</h2> <p>Our findings were quite disappointing. Among the 446 publications eligible for the analysis, QoL was apparently not listed among trial endpoints in 210 (47.1 %) (Fig. 1). The proportion of trials not including QoL among endpoints was surprisingly high (40.1 %), even among studies conducted in patients with advanced or metastatic setting, where the usually higher burden of symptoms and the balance between treatment efficacy (frequently modest) and toxicity should imply an increased attention to QoL. We divided trials into for-profit (when sponsored by a drug company) and no-profit (when sponsored by an academic institution or a cooperative group), and we found a high proportion of trials not including QoL among endpoints in both categories (39.7 % among for-profit trials and 53.6 % among non-profit trials).</p> <p><img src="/custom/img/files/files_datafiles_data_Zeitungen_2019_Leading Opinions_Onko_1904_Weblinks_lo_onko_1904_s50_fig1_dimaio.jpg" alt="" width="820" height="362" /></p> <h2>More attention to QoL data</h2> <p>Furthermore, our analysis showed that, despite QoL data are collected during treatment and their analysis does not need further follow-up, the inclusion of QoL among endpoints does not necessarily imply the presence of results in the primary study publication. Namely, out of 231 primary publications of trials with QoL as secondary or exploratory endpoint, QoL results were not available in 88 (38.1 %) (Fig. 2), with similar deficiency in trials conducted in patients with advanced or metastatic disease (37.6 %), in for-profit trials (37.1 %) and in non-profit trials (39.3 %). Overall, due to the absence of QoL among endpoints or to the missing publication of results, QoL data were not available in the majority of primary publications. Our analysis showed that attention to QoL is suboptimal in all the types of solid tumours, even in settings like lung cancer, where patients are often symptomatic and efficacy of treatments in terms of prolongation of survival is often disappointing. <br />If the publication of QoL results could be considered less essential for negative trials, which will not imply the subsequent adoption of experimental treatments in clinical practice, the absence of QoL data is particularly disappointing for trials with formally positive results according to the interpretation of primary endpoint. In fact, when the experimental treatment shows a benefit in terms of overall survival, QoL results are helpful to define the trade-off between benefits and harms, helping to define the value, especially if the survival benefit is modest and the potential toxicity of the treatment is not negligible. QoL results are even more important when the interpretation of the trial is based on a surrogate endpoint (such as progression-free survival): in this case, we strongly believe that a patient-focused assessment should be crucial to define the value of a radiologically defined result.</p> <p><img src="/custom/img/files/files_datafiles_data_Zeitungen_2019_Leading Opinions_Onko_1904_Weblinks_lo_onko_1904_s50_fig2_dimaio.jpg" alt="" width="820" height="397" /></p> <h2>Methodological difficulties to be faced</h2> <p>Of course, adoption of QoL among endpoints implies several methodological issues.² The choice of the correct questionnaires and the best timing of administration are crucial. Missing data can be a relevant problem, especially in diseases such as advanced lung cancer, where the proportion of early treatment failures is often not negligible. This can be a problem for the analysis, especially if the proportion of missing patients is unbalanced among study arms, considering that patients who do not fill in the questionnaires are mostly those who are getting worse. Furthermore, methodology of analysis and presentation of results can be really heterogeneous. Commonly, mean scores at different time points or mean changes compared to baseline are presented, but this analysis does not allow to understand how many patients experienced a significant improvement (or a significant worsening) of QoL, information better evaluable describing the proportion of responders, which is, however, presented only in a minority of papers. In addition, time to deterioration of specific symptoms or of global QoL is particularly useful to focus on treatment failure, helping to define if a radiologically defined benefit (like progression-free survival) is sustained by a subjective benefit. Whichever the modality of presentation of results, as for the other endpoints of treatment efficacy, when reading and interpreting QoL data, clinical relevance of the difference between arms should always be considered in addition to statistical significance. In addition to the presentation of results, all papers should explicit and reference the definition of minimum clinically important difference for all the QoL scores included in the analysis. <br />Although we are aware of the above described methodological difficulties of QoL analysis, we believe that cancer community should strongly require the adoption and timely reporting of patient-reported outcomes in clinical trials. All stakeholders (trial sponsors, clinical researchers and methodologists, cancer patients, regulatory agencies, scientific journals) should agree for inclusion of QoL among study endpoints and for the completeness of reporting of results in scientific publications.</p></p> <p class="article-footer"> <a class="literatur" data-toggle="collapse" href="#collapseLiteratur" aria-expanded="false" aria-controls="collapseLiteratur" >Literatur</a> <div class="collapse" id="collapseLiteratur"> <p><strong>1</strong> Marandino L et al.: Deficiencies in health-related qualityof- life assessment and reporting: a systematic review of oncology randomized phase III trials published between 2012 and 2016. Ann Oncol 2018; 29: 2288-95 <strong>2</strong> Fallowfield LJ: Quality of life assessment using patient-reported outcome (PRO) measures: still a Cinderella outcome? Ann Oncol 2018; 29: 2286-7</p> </div> </p>