Delays in Breast Cancer Diagnosis and Treatment: Influence on Survival
May 1, 1999
Delays in Breast Cancer Diagnosis and Treatment: Influence on Survival
abstracts & commentary
Synopsis: A new policy in the United Kingdom mandates the rapid referral of patients with breast symptoms for diagnosis and treatment. In a recent volume of Lancet, three articles and an editorial address the question of the influence of delay in treatment on survival.
Sources: Sainsbury R, et al. Lancet 1999;353: 1132-1135; Richards MA, et al. Lancet 1999;353: 1119-1126; Ramirez AJ, et al. Lancet 1999;353: 1127-1131; Coates AS, et al. Lancet 1999;353: 1112-1113.
It may be considered conventional wisdom that delayed presentation and treatment of cancer in general, and specifically breast cancer, is associated with lower survival.1 Thus in the UK, as of April 1, 1999, family physicians are required to refer all patients who have suspected breast cancer to the hospital within 14 days. This, of course, will reduce only one component of the normal delay. The other major component is patient delay (the time from first noticed symptoms to the seeking of medical help). In a recent issue of Lancet, three original articles and a commentary were published on this topic. Two papers examined the influence of delays on survival, with conflicting results, and the third described factors that are associated with delays.
Sainsbury and colleagues did a retrospective analysis of 36,222 patients with breast cancer listed in the Yorkshire Cancer Registry. They found that there was no evidence that provider delays of longer than 90 days adversely influenced survival. They found that delays from the first family physician visit to the initiation of treatment have not changed much in the past two decades. They did observe that younger patients were more likely to have delays. More than 8% of patients younger than 50 years delayed longer than 90 days, compared with 3% of patients older than 50 years. Similarly, 48% of younger patients had their first treatment within 30 days compared with 64% of those older than 50 years. Curiously, of the 5708 patients diagnosed between 1986 and 1990, there was a survival disadvantage for those who presented early and were treated in less than 30 days.
Richards and colleagues examined the question using a different method. They systematically reviewed 87 published reports of 101,954 patients that included direct data linking delay (including the delay of patients seeking medical attention) and survival. Reports were divided into three categories. In category I studies, actual five-year survival data was available, whereas in category II, actuarial or multivariate analyses were undertaken to estimate survival. Category III reports included all other types of data.
For the category I studies, patients with delays of three months or more had a 12% lower five-year survival than those with shorter delays (odds ratio for death 1.47 [95% CI 1.42-1.53]). In category II, 13 of 14 studies with unrestricted samples showed a significant adverse relation between longer delays and survival. In category III, all three studies with unrestricted patients had similar conclusions. However, in four of the five studies of only patients with operable disease, there was no association found between delay and survival. In the 13 studies in which the data were available, longer delays were associated with more advanced stage. In those that controlled for stage, longer delay was not associated with shorter survival when the effect of stage on survival was taken into account.
In the third paper, Ramirez and colleagues performed a systematic review of the large literature regarding risk factors for delay in obtaining breast cancer treatment. After an exhaustive review, Ramirez et al found that most studies were of insufficient quality to provide confident explanations for the delays. There was strong evidence of an association between older age and delay by patients and also strong evidence that marital status was unrelated to delays by patients. Younger age and presentation with a breast symptom other than a lump were strong risk factors for delays by providers.
COMMENTARY
Although the reports by Sainsbury et al would seem to conflict that of the Richards group, there may be a satisfactory explanation when the complexities of treatment delays are considered in the context of the methods used. As Coates explains in his commentary in the same Lancet issue, delay in treatment may be divided into three phases and these reports were examining different components of the delay. Delay by the patient is the interval between time of first symptom to the first medical consultation. Data regarding this delay may be confounded by the imprecise definition of early symptoms, and possible under reporting by patients. Provider delay (that examined by Sainsbury et al) covers the period from first consultation to referral. The third delay is that from referral to treatment. Reducing this third component is the subject of the recently initiated policy in the UK but no data supporting this mandate could be derived from that presented in these articles.
Two types of bias might influence the association of delay in treatment and survival. Lead-time bias arises when survival is measured from the time of diagnosis. Thus, early diagnosis might appear to be associated with prolonged survival even in the absence of effective therapy or, in contrast, delay in diagnosis might appear to be associated with shortened survival. This type of bias might be partially overcome if survival is measured from the time of first symptoms, but even this is fraught with imprecision. In one previously reported study, survival was measured for breast cancer patients from the time of first symptom and it was adversely affected by the delay from symptom appearance to diagnosis.2
A second type of bias relates to the rate of tumor growth, symptom development, diagnosis, and treatment. A rapidly growing tumor might result in prompt patient presentation and a sense of urgency on the part of providers to initiate treatment. Under this circumstance, a shorter interval would be recorded from presentation to diagnosis to treatment (i.e., short delay) and a shorter survival (due to the nature of aggressive tumors) observed. Thus, for patients with more aggressive disease it might appear that shorter delay to diagnosis and treatment is associated with shorter survival.
Thus, with these (and other) inherent complexities in determining the importance of delay in diagnosis and treatment for breast cancer, common sense must prevail. As Coates indicates, breast cancer is not an emergency. A careful and systematic pretreatment evaluation and multidisciplinary treatment approach would seem preferable to a haphazard rush to accomplish treatment with no delay. Certainly, efforts to reduce the component of delay from first symptom to medical consultation are worthy of investigation and implementation.
References
1. Caplan LS, et al. Public Health Rev 1993;20:187-214.
2. Elwood JM, et al. BMJ 1980;280:1291-1294.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content