Primary Care Physicians and Oncologists Are Players on the Same Team

The increasing incidence and prevalence of cancer, the growing number and complexity of cancer therapies, and the emerging recognition of the unique aspects of cancer survivorship are factors that remind us again and again that it takes a health care team to care for patients with cancer. Primary care physicians (PCPs) are crucial to that team, and numerous studies have shown that they are keen to be team players for their patients with cancer. Certainly, the concept of shared care as the key to optimizing the benefits of both primary care and specialist care has long been promoted as the way forward for patients with cancer. The shared management approach during active cancer treatment has been studied previously through standardized information transmitted to community-based PCPs with specialist oncology care establishing the treatment protocol. However, PCPs frequently need a little coaching to help them be most effective when providing care for their patients undergoing active cancer treatments. In this issue, Jefford et al report the results of a randomized controlled trial designed to improve the information about chemotherapy regimens that is sent to PCPs with patients on active treatment. The intervention consisted of specifically developed information sheets about the chemotherapy regimen and associated adverse effects. The regimen-specific information sheets were transmitted by fax together with a covering letter, and the primary care practice was telephoned to ensure the fax was received and filed in the patient’s chart. This intervention fulfills several of the important principles of effective knowledge translation: the information was tailored to an individual patient, specific to primary care in terms of content and format, and sent in a timely manner to facilitate optimal management of the patient’s care. Although this intervention was not multifaceted, another important aspect of effective knowledge translation, the design, was elegant and pragmatic, which enhances its generalizability. The strength of the intervention is its relative simplicity. It would appear, however, that calling the practice and ensuring the information was placed on the patient’s chart is actually an important component: 17 (39%) of 44 PCPs in the control group reported that they did not receive any information as compared with only three (8%) of 37 PCPs in the intervention group. Even with this extra step, it is difficult to imagine that this simple intervention could not be implemented in most, if not all, cancer treatment facilities. Although a formal economic evaluation was not conducted, the simplicity of the intervention strongly suggests that it would be cost effective. Despite the presumed low cost, however, for an intervention to be cost effective, it must be shown to be effective. In this trial, the outcomes studied were PCPs’ confidence in managing patients receiving chemotherapy, PCPs’ knowledge of indications for referral and chemotherapy adverse effects that their patient might experience, and PCPs’ satisfaction with the information received and with shared care. The intervention was effective in terms of statistical significance for the outcomes of confidence and satisfaction, yet the clinical importance of these improvements is not known. The knowledge outcome did not improve. A substantial concern is that the instruments used to measure these outcomes were not rigorously developed, and therefore one cannot be confident that they are valid measures of the constructs being studied. While the process for developing the intervention was appropriate, the development of the outcome measures did not meet the same standard. Consider, for example, the instrument used to measure satisfaction. The instrument consisted of two items: one item measured satisfaction with shared care, and one item measured satisfaction with communication. The construct of satisfaction is complex and multidimensional, as illustrated by the intricate development processes and multi-item scales used to measure patient satisfaction. One suspects that like patient satisfaction, PCP satisfaction is equally multidimensional, requiring more than one item on a two-item scale to measure each dimension. Similarly, as noted by the authors, the failure to detect a difference in knowledge might have been due to the limitations of the measurement instrument. The strong and influential methodology of a randomized controlled trial, rigorously applied as it was in this study, is weakened if the validity of the instruments used to measure outcomes is questionable. The study would have been markedly strengthened if the thorough process used to develop the intervention had similarly been used to develop and test the outcome measures before starting the trial. A matter of further concern is whether these measures—if we accept that they are valid measures of knowledge, confidence, and satisfaction—actually lead to better patient care and, subsequently, improved patient outcomes. Such a relationship is neither apparent nor inevitable. While the outcomes of PCP confidence and satisfaction may or may not be necessary, they are certainly not sufficient to ensure improvements in patient care. A vital next stage of this research, therefore, would be to measure the effectiveness of the intervention on patient outcomes such as symptom management, emergency room visits, unscheduled visits to cancer specialists, and patient satisfaction. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 26 NUMBER 14 MAY 1