Identifying key symptoms

The QL assessment typically generates huge amounts of data, and it is therefore important to identify a primary QL outcome, and a few secondary outcomes. Although further analyses are always possible, if they have not been pre-defined they should always be considered as hypothesis generating.

Sometimes a new treatment is specifically aimed at one symptom, and the primary QL outcome can relate to this. For instance, Sur et al. [37] tested the effect of sucalfate to palliate radiation oesophagitis, and Marty et al. [38] assessed granisetron as an anti-emetic in cytostatic-induced emesis. In these situations it is clear that the severity of the key symptom before, possibly during, and after treatment needed to be assessed. In addition it may be useful to assess the effect of these changes on other aspects of patients' lives. For instance, if oesophagitis improves, how does this affect appetite or social functioning?

It may also be important not to focus only on the possible positive aspects of a treatment. Treatment-related adverse effects may also be relevant. For example, in a trial of two thoracic radiotherapy regimens for lung cancer, the key QL outcome might be the duration and severity of dysphagia. In an MRC Lung Cancer Working Party trial [39], patients with non-small cell lung cancer were asked to complete a daily diary card, and the proportion of patients reporting moderate or severe dysphagia was plotted. This suggested that in terms of dysphagia, the shorter radiotherapy schedule (17 Gy in 2f) affected fewer patients and was transient, whereas the longer radiotherapy schedule affected more patients for longer (Fig. 6.3).

Proportion 100-of

17 Gy/2f 39 Gy/13f patients 80-reporting moderate 60-or severe dysphagia 40-

moderate 60-or severe dysphagia 40-

7 14 21

Days from start of radiotherapy

Fig. 6.3 Proportion of patients reporting moderate or severe dyspagia on a daily diary card (adapted from MRC Lung Cancer Working Party, [39]).

Sometimes more than one symptom may be important. For instance, patients may present with a complex mixture of symptoms which the treatment should palliate. In this situation a combination score or an algorithm maybe considered. In a trial assessing the value of mitoxantrone and prednisone in twenty-seven patients with hormonally resistant prostate cancer, Moore etal. [40] pre-defined a palliative response as: a decrease in analgesic score by >50 per cent or a decrease in 'present pain intensity' by >2 points without an increase in analgesic score. In this phase II study nine patients were considered 'palliated' using this trial-specific definition, compared with only one who showed a traditional partial response. In an MRC Lung Cancer Working Party trial [41] comparing oral chemotherapy versus standard intravenous chemotherapy in patients with small cell lung cancer, QL was considered to be a primary outcome. In order to be considered 'equivalent' the oral treatment was required 'to achieve at least equivalent palliation of major symptoms (cough, pain, anorexia and shortness of breath) at three months from randomization.' The four scores for these symptoms were added together to get a score at baseline and at three months. Palliation was defined as a reduction in the sum of the scores from baseline to three months (indicating an improvement).

When planning a trial it is crucial to decide what change in QL would represent worthwhile benefit. Burris and Storniolo [42], in a trial of two chemotherapy treatments for pancreatic cancer, created an algorithm (Fig. 6.4) to define worthwhile benefit and to divide their patients into QL responders and non-responders. The algorithm combined pain scores and performance status (PS), with weight change as a tiebreaker if pain and PS were stable. So far, few other groups have used this approach, even though there

Fig. 6.4 An algorithm to define clinical benefit in a trial of pancreatic cancer (adapted from Burris and Storniolo, [42]).

would be strong arguments for using it in future trials of pancreatic cancer, and for developing standard algorithms for other cancers. Such an approach would also permit cross-trial comparisons.

Before formulating a QL hypothesis, a systematic search of the literature should be performed. This should reveal if any work in that relevant area could be used, as it is generally better to repeat and develop ideas rather than unnecessarily start from the beginning again. However, as so little work has been performed in the field ofdeveloping relevant QL hypotheses, the choice of key symptoms or the development of an algorithm will often be unique to the trial.

It would seem that with careful background work in most trials a hypothesis can be formed which directs the QL assessment. This does not mean that other outcomes should be ignored (this situation is analogous to powering a trial for duration of survival but still evaluating response and toxicity), but it does mean that the focus can be on the key QL issues.

Quit Smoking Today

Quit Smoking Today

Quit smoking for good! Stop your bad habits for good, learn to cope with the addiction of cigarettes and how to curb cravings and begin a new life. You will never again have to leave a meeting and find a place outside to smoke, losing valuable time. This is the key to your freedom from addiction, take the first step!

Get My Free Ebook


Post a comment