Tools Help Prostate Cancer Patients Participate in Decision

With no clearly superior treatment for localized prostate cancer, physicians and patients would like to increase patient participation in the decision-making process. Unfortunately, physicians frequently have difficulty understanding patients’ preferences, and patients often do not have sufficient knowledge to make an informed treatment decision. Shared- decision-making tools, such as decision analyses, may increase patient participation and thereby improve physicians’ understanding of their patients’ views.

Different ideas about the objectives of treatment and miscommunication about treatment issues often characterize patient-physician interactions. In a survey of prostate cancer patients, Crawford demonstrated that patients and physicians differ in what they believe should be the goals of treatment. While patients viewed preservation of quality of life and extension of life as the most important treatment goals, urologists focused on treatment effectiveness as the most important consideration.[1]

Patients and physicians also differed in their descriptions of the patient-physician encounter. Only one fifth of patients remembered having discussed their preferences, the costs of treatment, and the side effects of prostate cancer and its treatment with their physician. In comparison, all urologists queried recalled discussing these issues, including alternative treatment options.[1]

In another study of patient-physician interactions, physicians could not specify their patients’ concerns, even when patients remembered that their doctor asked for this information.[2]

These studies suggest that physician understanding of patient preferences and patient participation in treatment decision making need to be addressed.

Shared-decision-making tools, including CD-ROMs, videotapes, brochures, and decision analyses, can educate patients and can aid doctors in understanding patients’ preferences and including them in the decision-making process. Unfortunately, many of these tools require constant updates of the material presented, apply group-level recommendations to individual patients, and do not educate patients well enough to allow them to make informed decisions.[3]

Tools such as decision analyses that can easily incorporate the most recent research data, individual clinical characteristics, and patient preferences may facilitate decision making.

VA Study

A recent study evaluated a shared-decision-making tool in 13 newly diagnosed prostate cancer patients during their initial meeting with the treating urologist at the VA Chicago Health Care System-Lakeside Division. This study employed a computer-based decision analytic model that combines patient preferences, pathologic characteristics of the cancer, patient age, and patient co-morbidities, to generate individualized treatment recommendations for surgery, radiation therapy, and watchful waiting.

This study examined the consistency of the patients’ preferences, the concordance between the decision analytic model recommendation and the physician recommendation, and possible factors that contributed to the physician’s recommendation but were not taken into account by the model.

Newly diagnosed localized prostate cancer patients were administered the interviewer-assisted computer-based decision-analytic model.[4,5] The model measured patients’ preferences for
potential health states (see Table) associated with prostate cancer and its treatment—including impotence, incontinence, watchful waiting, combined impotence and incontinence, and current health—and integrated these preferences into the model to produce quality-adjusted life expectancies (QALEs).

Patients’ preferences for each health state were evaluated by identifying the risk each person would take in order to avoid that health state.

For the health states of impotence, incontinence, and combined impotence and incontinence, the patient was asked to choose between two treatment op-tions: (1) one that cured the patient’s prostate cancer but resulted in the patient having impotence, incontinence, or both; and (2) one that promised to cure the patient’s prostate cancer without impotence, incontinence, or both, but carried a certain percent chance of death. The treatment options were presented with different chances of death until the patient could not decide on either possible outcome. Graphic displays were used to aid patients with low-literacy levels.

The decision analytic model produced first, second, and third best treatment options based on each patient’s preferences, pathologic status, and clinical characteristics. After examining and reviewing a patient’s records, the physician recommended his own treatment. Interviews were also conducted with the physician to learn what factors he considered in making his decision.

Of the 13 patients, 9 were white and 4 were black. The mean age was 68.9 years. Only one patient did not finish high school, 6 had some college education, and 2 had received college degrees.

Study Results

Four of the 13 patients chose not to take any risk to avoid the potential health states associated with prostate cancer and its treatment. One patient was only willing to take a risk to avoid the health state of watchful waiting (in which the patient chose between treating his prostate cancer and "watchful waiting") but not for the other potential health states.

Seven patients had inconsistent preferences, rating the health state of combined sexual and bladder dysfunction higher, or better, than either impotence or incontinence alone. These 7 patients were more willing to take risks to avoid the health state of either impotence or incontinence than the combined health state of impotence and incontinence.

For 6 of the 13 patients, the physician recommendation did not coincide with the model recommendation. For one of the discrepancies, the physician recommended surgery over the model’s proposal of watchful waiting. In this case, the model’s recommendation of watchful waiting had a 1 year greater adjusted life expectancy than the physician’s recommendation of surgery.

In another discordant case, the model recommended surgery compared with the physician’s recommendation of radiation therapy. The model’s recommendation came with a QALE advantage of three tenths of a year over the physician’s recommendation.

In the remaining 4 discordant recommendations, all of which occurred in older patients (70 years of age or older), the physician recommended watchful waiting, compared with the model’s recommendation of radiation therapy. The differences in quality-adjusted life expectancy for these patients ranged from 0.04 to 1.2 years. In each case, older age and psychosocial concerns appeared to influence the urologist.

Discussion

Shared-decision-making tools may improve patient participation and physician understanding of patients’ preferences, and decision analysis models may offer a valid and feasible approach to shared decision-making. The major concern with this decision analytic approach appears to be the unexpected preferences given by the patients: 7 of the 13 patients had inconsistent preferences, and 4 were unwilling to take any risks at all to avoid the potential health states associated with prostate cancer and its treatment.

There are several possible explanations for the patients’ unexpected responses. First, the patients who did not wish to take any risks to avoid the health states may not view taking a gamble with their life as a rational choice. Second, the patient who was only willing to take a risk to avoid not pursuing any treatment at all (watchful waiting) may have responded in this manner due to his anxiety about living with untreated cancer.

Several patients rated the combined health state of bladder and sexual dysfunction higher than either dysfunction alone. This inconsistency may have been due to patients experiencing a "learning curve" and reevaluating their preferences during the assessment. Their later preferences may in fact be more accurate.

Finally, although attempts were made to extract patients’ preferences in an easily understandable, implicit way, participant education level may also have hindered understanding of these tasks, thus contributing to such unexpected scores.

It is interesting that the model recommended radiation therapy more than the physician, especially in older patients. This may be due to the clinical inputs related to the effectiveness of the treatment used in the model to generate QALEs for watchful waiting, radiation therapy, and surgery. The large differences in life expectancy between model and physician recommendation may be due to the less conservative mentality of the model or the fact that the physician probably considers many other factors such as cost or burden of travel.

There are two limitations of this study. First, two commonly used treatments, neoadjuvant hormonal therapy and brachytherapy, were not considered as treatment options in the model because of lack of data on complication rates for hormonal therapy and infrequent use of the treatment in the clinical setting for brachytherapy. Second, recent studies show that prostate cancer treatment complications are affected by time. Thus, an evaluation of a decision analytic model with more realistic scenarios including time course of complications is needed.

Whether this decision analysis model is a reliable decision-making tool is still uncertain. Improved methods of assessing patient preferences may enhance the usability of this decision model. Overall, this study revealed the need for further development of decision analytic models for improving the participation of newly diagnosed, localized prostate cancer patients in treatment decision-making. 

References:

1. Crawford DE et al: Comparison of perspectives on prostate cancer: Analysis of survey data. Urology 50:336, 1997.

2. Knight SJ et al: Perceived involvement with care, patient satisfaction, and patient-clinician agreement about prostate cancer utilities. ASCO 35th Annual Meeting, Atlanta, 1999, abstract 1589.

3. Kaplan S: The future of patient input into medical decision making (editorial). Quality Review Bulletin 18:182, 1992.

4. Kattan MW et al: A decision analysis for treatment of clinically localized prostate cancer. J Gen Intern Med 12: 299, 1997.

5. Beck JR et al: A critique of the decision analysis for clinically localized prostate cancer. J Urol 152:1894, 1994.