A Pilot Trial of Decision Aids to Give Truthful Prognostic and Treatment Information to Chemotherapy Patients with Advanced Cancer

Primary Outcome

Our primary outcome was to assess if patients would complete a decision aid with full disclosure. Of 27 patients, only one (4%) chose not to complete the decision aid after starting. She was a 55-year-old African American woman who had recently started first-line treatment for metastatic colorectal cancer. She had been told at another institution that she had lost too much weight and was too ill to benefit from chemotherapy, but with counseling she regained the weight and had a performance status of 2 at VCU. In her pretest, she answered that she thought a woman with metastatic colorectal cancer spread to bones and lymph glands could be cured, with a chance of cure of 50%. Once presented with the information (good treatments that prolong life and control symptoms but no chance of cure and 9% of patients with metastatic colorectal cancer alive at 5 years), she said that she did not want to finish the questions. She did complete her HHI, which did not change, and was not distressed (see Table 2, patient 12).

In the pretest, almost all the patients, including the patient above, reported wanting full disclosure about cancer, prognosis, treatment, and side effects. In response to questions beginning “How much do you want to know about …” 27 of 27 answered “Tell me all” to the questions about “your cancer,” “your prognosis,” “treatment benefits,” and “treatment side effects.” Only one of 27 answered otherwise: “Tell me a little” about cancer, and “Tell me some” about prognosis.

Secondary Outcomes

Participants were overoptimistic about the results of palliative chemotherapy, as shown in Table 3. Most (14/27, 52%) people thought a person with “metastatic cancer (breast, colorectal, lung, prostate—specific to that person's disease) spread to the bones and lymph glands” could be cured. After the decision aid, more people recognized that their cancer could not be cured (17/25, 63%) but eight of 25 (32%, P = 0.15, Fischer's exact test) still thought a person with metastatic disease could be cured. Patients were particularly overoptimistic about the chance of their symptoms being helped by chemotherapy: 87% thought their symptoms would be helped by chemotherapy, and 60% thought a patient would have at least 50% shrinkage of their cancer before the exercise, which declined only slightly after the decision aid. (While the correct answer varies by disease, the number helped by chemotherapy is usually less than 50%, and response rates are always less than 50%.)

There was no change in responses to the HHI after the intervention as we have previously reported.¹⁸ Participants did not appear to be visibly distressed by the intervention. A psychologist and chaplain were made available, but no one requested their services. In our small clinic, the primary nurses and doctors have frequent interactions during visits and chemotherapy. No patient was reported to be distressed in any way, during that visit or subsequent visits.

The comments recorded by the patients or the interviewers at the end of the exercise showed that most patients would share the information, as shown in Table 4.

In some cases, the average prognosis and treatment benefit, although small, was bigger than the person thought before the exercise. Nearly all found it helpful. Some illustrative comments are shown in Table 2.

We did not formally measure the time to complete the screening questions, pre- and posttests, pre- and post-HHI, and decision aid; but in most cases it took less than 20 minutes to complete the whole package including the pre- and post-tests. Review of the decision aid with the patient always took less than 5 minutes, even when we were reading it with the patient and family. This is consistent with work showing that oncologists state that completing an advance directive will take too much time but, in fact, it takes less than 10 minutes.^{[19] and [20]}

Discussion

Historical data show that patients know little about their prognosis and the effect that treatment will have on their cancer. Yet, this knowledge is essential to making informed choices about treatment benefits, risks, and even costs. When tested in randomized controlled trials, decision aids led to more involvement in decision making.^{[21] and [22]} However, there were no decision aids available about metastatic incurable disease, despite some promising early starts^{[23], [24], [25], [26], [27] and [28]} and only one about first-line treatment,²⁹ so we made a simple one. A successful decision aid may allow patients to discuss their situations with their physicians and develop management strategies that best concur with personal goals and preferences and help patients make plans in other areas of life.

Our findings suggest that most people do want honest information, even if the news is bad. We found that 27 of 27 enrolled patients initially reported wanting to know all the available information about their cancer, prognosis, treatment benefits, and treatment side effects. Also, 26 of 27 patients were able to complete the decision aid fully, our main outcome measure. While approximately 10% of available patients were excluded from accrual by their oncologists or oncology nurses due to preexisting distress, fear of distress in the patient or family member, uncontrolled symptoms, or psychiatric illness, in general there was excellent acceptance of the study by patients and oncologists. In this pilot study we did not investigate the attitudes of nonparticipants nor were we able to collect sociodemographic data to determine nonresponse bias, that is, whether certain types of patients are more likely to decline participation in the study.

Participants in the study were overoptimistic about their chances of cure, potential treatment response, symptom relief, and survival. None of these patients had curable disease, but 63% thought that a person with metastatic cancer of their type could be cured and gave the average chance of cure as 52%. Inaccurate assessment of cure rates decreased postintervention. At the pretest 14/27 (52%) believed a person with cancer similar to theirs could be cured, which changed to 8/26 (31%) at the posttest. This agrees with other studies that showed that patients mistook palliative radiation for curative radiation about one-third of the time, even when provided with accurate information.^{[1], [30] and [31]}

Knowledge of prognosis and planning for the future is important as there is evidence of benefit to having the discussion about treatment outcomes. Recent data show improved quality of care, improved quality of life, and improved caregiver quality of life if the physician discusses death with the patient and family.⁵ Transplant patients with advanced directives had more than a twofold survival advantage over those without them.²⁷ Conversely, over- or underestimating survival or treatment benefit can lead to bad health outcomes. Stem-cell transplant patients who were overoptimistic lived no longer than those with realistic views.^{[32], [33] and [34]} Cancer patients who overestimated their survival were more likely to die a “bad” death (defined as death in an intensive care unit, on a ventilator, or with multiple hospitalizations and emergency room visits) without achieving life extension.³⁵ It may be that the 16%–20% of patients with incurable solid tumors who start a new chemotherapy regimen within 2 weeks of death,³⁶ when they are unlikely to benefit, simply do not know the prognosis or treatment effect or have different perspectives.³⁷ Alternatively, we do not know how many patients decline second- and nth-line chemotherapy without knowing the full benefits and risks and who might choose chemotherapy if they knew second- or nth-line chemotherapy improved survival, pain scores, or quality of life. For instance, 40% of breast cancer patients will have some disease control from fourth-line chemotherapy for up to 4 months even if there is no evidence of improved survival.³⁸

Patients consistently tell us to be truthful, compassionate, and clear and to stay the course with them.^{[39] and [40]} Despite nearly all American patients stating that they want full disclosure about their prognosis, treatment options, and expected outcomes, most patients do not receive such information⁴¹ or receive such information far too late in their course.⁴² Even if terminally ill patients with cancer requested survival estimates, doctors would provide such estimates only 37% of the time, often an overestimate;⁷ and a recent meta-analysis showed that cancer physicians consistently overestimated prognosis by at least 30%.⁴³ Honest information respects the autonomy of a patient to make decisions based on what is known about the outcomes of such decisions.⁴⁴ Such information should not be forced on a patient, but the patient should be told that the information is available and that he or she has the right to accept or decline the information.⁴⁵

When we started this project, colleagues were concerned about whether patients would want such information, that patients would be distressed by poor prognosis, that patients would give up hope, and that the procedures would take too much time. We also were concerned about the effect of giving such bad news on the provider, when prior research showed negative effects on the information-giver's mood and affect from such encounters⁴⁶ and that doctors in general protect themselves by not giving bad news.⁴⁷ Completion of the decision aid was difficult for the interviewers, too. Some commented on how hard it was to give “bad” information about chance of cure and expected survival, even for patients they did not know. While patients may be more comfortable having advance directive discussions with a doctor they do not know rather than their oncologist,⁴⁸ it can still be hard for the provider. Surprisingly, it rarely took more than 20 minutes to discuss the information including the tests since the information was preprinted.

Patients vary in their approach to decision making, but the decisions should at least start with good information. Based on these preliminary findings, the piloted intervention is significant because it can lead to measurable impacts on knowledge about prognosis and appears to be judged helpful. We do not know the impact of full and truthful information on patient knowledge, decision making, hope, attendant choices about advanced medical directives, chemotherapy use, or hospice use. The next steps are to make the information available directly to patients on the Internet, which is in progress. The purpose is not to increase or decrease the use of palliative chemotherapy or hospice care; the lack of research into the decisions fully informed patients make precludes any such prediction. Since the intervention appears to be successful in this pilot trial, it will be tested in conjunction with standard care in a randomized clinical trial with measurement of quality of care, quality of life, and health-care cost outcomes.

Acknowledgments

This research was supported by VCU School of Medicine Research Year Out, GO8 LM0095259 from the National Library of Medicine (T. J. S., L. L., J. K.), and R01CA116227-01 (T. J. S.) from the National Cancer Institute.

Primary Outcome

Our primary outcome was to assess if patients would complete a decision aid with full disclosure. Of 27 patients, only one (4%) chose not to complete the decision aid after starting. She was a 55-year-old African American woman who had recently started first-line treatment for metastatic colorectal cancer. She had been told at another institution that she had lost too much weight and was too ill to benefit from chemotherapy, but with counseling she regained the weight and had a performance status of 2 at VCU. In her pretest, she answered that she thought a woman with metastatic colorectal cancer spread to bones and lymph glands could be cured, with a chance of cure of 50%. Once presented with the information (good treatments that prolong life and control symptoms but no chance of cure and 9% of patients with metastatic colorectal cancer alive at 5 years), she said that she did not want to finish the questions. She did complete her HHI, which did not change, and was not distressed (see Table 2, patient 12).

In the pretest, almost all the patients, including the patient above, reported wanting full disclosure about cancer, prognosis, treatment, and side effects. In response to questions beginning “How much do you want to know about …” 27 of 27 answered “Tell me all” to the questions about “your cancer,” “your prognosis,” “treatment benefits,” and “treatment side effects.” Only one of 27 answered otherwise: “Tell me a little” about cancer, and “Tell me some” about prognosis.

Secondary Outcomes

Participants were overoptimistic about the results of palliative chemotherapy, as shown in Table 3. Most (14/27, 52%) people thought a person with “metastatic cancer (breast, colorectal, lung, prostate—specific to that person's disease) spread to the bones and lymph glands” could be cured. After the decision aid, more people recognized that their cancer could not be cured (17/25, 63%) but eight of 25 (32%, P = 0.15, Fischer's exact test) still thought a person with metastatic disease could be cured. Patients were particularly overoptimistic about the chance of their symptoms being helped by chemotherapy: 87% thought their symptoms would be helped by chemotherapy, and 60% thought a patient would have at least 50% shrinkage of their cancer before the exercise, which declined only slightly after the decision aid. (While the correct answer varies by disease, the number helped by chemotherapy is usually less than 50%, and response rates are always less than 50%.)

There was no change in responses to the HHI after the intervention as we have previously reported.¹⁸ Participants did not appear to be visibly distressed by the intervention. A psychologist and chaplain were made available, but no one requested their services. In our small clinic, the primary nurses and doctors have frequent interactions during visits and chemotherapy. No patient was reported to be distressed in any way, during that visit or subsequent visits.

The comments recorded by the patients or the interviewers at the end of the exercise showed that most patients would share the information, as shown in Table 4.

In some cases, the average prognosis and treatment benefit, although small, was bigger than the person thought before the exercise. Nearly all found it helpful. Some illustrative comments are shown in Table 2.

We did not formally measure the time to complete the screening questions, pre- and posttests, pre- and post-HHI, and decision aid; but in most cases it took less than 20 minutes to complete the whole package including the pre- and post-tests. Review of the decision aid with the patient always took less than 5 minutes, even when we were reading it with the patient and family. This is consistent with work showing that oncologists state that completing an advance directive will take too much time but, in fact, it takes less than 10 minutes.^{[19] and [20]}

Discussion

Historical data show that patients know little about their prognosis and the effect that treatment will have on their cancer. Yet, this knowledge is essential to making informed choices about treatment benefits, risks, and even costs. When tested in randomized controlled trials, decision aids led to more involvement in decision making.^{[21] and [22]} However, there were no decision aids available about metastatic incurable disease, despite some promising early starts^{[23], [24], [25], [26], [27] and [28]} and only one about first-line treatment,²⁹ so we made a simple one. A successful decision aid may allow patients to discuss their situations with their physicians and develop management strategies that best concur with personal goals and preferences and help patients make plans in other areas of life.

Our findings suggest that most people do want honest information, even if the news is bad. We found that 27 of 27 enrolled patients initially reported wanting to know all the available information about their cancer, prognosis, treatment benefits, and treatment side effects. Also, 26 of 27 patients were able to complete the decision aid fully, our main outcome measure. While approximately 10% of available patients were excluded from accrual by their oncologists or oncology nurses due to preexisting distress, fear of distress in the patient or family member, uncontrolled symptoms, or psychiatric illness, in general there was excellent acceptance of the study by patients and oncologists. In this pilot study we did not investigate the attitudes of nonparticipants nor were we able to collect sociodemographic data to determine nonresponse bias, that is, whether certain types of patients are more likely to decline participation in the study.

Participants in the study were overoptimistic about their chances of cure, potential treatment response, symptom relief, and survival. None of these patients had curable disease, but 63% thought that a person with metastatic cancer of their type could be cured and gave the average chance of cure as 52%. Inaccurate assessment of cure rates decreased postintervention. At the pretest 14/27 (52%) believed a person with cancer similar to theirs could be cured, which changed to 8/26 (31%) at the posttest. This agrees with other studies that showed that patients mistook palliative radiation for curative radiation about one-third of the time, even when provided with accurate information.^{[1], [30] and [31]}

Knowledge of prognosis and planning for the future is important as there is evidence of benefit to having the discussion about treatment outcomes. Recent data show improved quality of care, improved quality of life, and improved caregiver quality of life if the physician discusses death with the patient and family.⁵ Transplant patients with advanced directives had more than a twofold survival advantage over those without them.²⁷ Conversely, over- or underestimating survival or treatment benefit can lead to bad health outcomes. Stem-cell transplant patients who were overoptimistic lived no longer than those with realistic views.^{[32], [33] and [34]} Cancer patients who overestimated their survival were more likely to die a “bad” death (defined as death in an intensive care unit, on a ventilator, or with multiple hospitalizations and emergency room visits) without achieving life extension.³⁵ It may be that the 16%–20% of patients with incurable solid tumors who start a new chemotherapy regimen within 2 weeks of death,³⁶ when they are unlikely to benefit, simply do not know the prognosis or treatment effect or have different perspectives.³⁷ Alternatively, we do not know how many patients decline second- and nth-line chemotherapy without knowing the full benefits and risks and who might choose chemotherapy if they knew second- or nth-line chemotherapy improved survival, pain scores, or quality of life. For instance, 40% of breast cancer patients will have some disease control from fourth-line chemotherapy for up to 4 months even if there is no evidence of improved survival.³⁸

Patients consistently tell us to be truthful, compassionate, and clear and to stay the course with them.^{[39] and [40]} Despite nearly all American patients stating that they want full disclosure about their prognosis, treatment options, and expected outcomes, most patients do not receive such information⁴¹ or receive such information far too late in their course.⁴² Even if terminally ill patients with cancer requested survival estimates, doctors would provide such estimates only 37% of the time, often an overestimate;⁷ and a recent meta-analysis showed that cancer physicians consistently overestimated prognosis by at least 30%.⁴³ Honest information respects the autonomy of a patient to make decisions based on what is known about the outcomes of such decisions.⁴⁴ Such information should not be forced on a patient, but the patient should be told that the information is available and that he or she has the right to accept or decline the information.⁴⁵

When we started this project, colleagues were concerned about whether patients would want such information, that patients would be distressed by poor prognosis, that patients would give up hope, and that the procedures would take too much time. We also were concerned about the effect of giving such bad news on the provider, when prior research showed negative effects on the information-giver's mood and affect from such encounters⁴⁶ and that doctors in general protect themselves by not giving bad news.⁴⁷ Completion of the decision aid was difficult for the interviewers, too. Some commented on how hard it was to give “bad” information about chance of cure and expected survival, even for patients they did not know. While patients may be more comfortable having advance directive discussions with a doctor they do not know rather than their oncologist,⁴⁸ it can still be hard for the provider. Surprisingly, it rarely took more than 20 minutes to discuss the information including the tests since the information was preprinted.

Patients vary in their approach to decision making, but the decisions should at least start with good information. Based on these preliminary findings, the piloted intervention is significant because it can lead to measurable impacts on knowledge about prognosis and appears to be judged helpful. We do not know the impact of full and truthful information on patient knowledge, decision making, hope, attendant choices about advanced medical directives, chemotherapy use, or hospice use. The next steps are to make the information available directly to patients on the Internet, which is in progress. The purpose is not to increase or decrease the use of palliative chemotherapy or hospice care; the lack of research into the decisions fully informed patients make precludes any such prediction. Since the intervention appears to be successful in this pilot trial, it will be tested in conjunction with standard care in a randomized clinical trial with measurement of quality of care, quality of life, and health-care cost outcomes.

Acknowledgments

This research was supported by VCU School of Medicine Research Year Out, GO8 LM0095259 from the National Library of Medicine (T. J. S., L. L., J. K.), and R01CA116227-01 (T. J. S.) from the National Cancer Institute.

Primary Outcome

Our primary outcome was to assess if patients would complete a decision aid with full disclosure. Of 27 patients, only one (4%) chose not to complete the decision aid after starting. She was a 55-year-old African American woman who had recently started first-line treatment for metastatic colorectal cancer. She had been told at another institution that she had lost too much weight and was too ill to benefit from chemotherapy, but with counseling she regained the weight and had a performance status of 2 at VCU. In her pretest, she answered that she thought a woman with metastatic colorectal cancer spread to bones and lymph glands could be cured, with a chance of cure of 50%. Once presented with the information (good treatments that prolong life and control symptoms but no chance of cure and 9% of patients with metastatic colorectal cancer alive at 5 years), she said that she did not want to finish the questions. She did complete her HHI, which did not change, and was not distressed (see Table 2, patient 12).

In the pretest, almost all the patients, including the patient above, reported wanting full disclosure about cancer, prognosis, treatment, and side effects. In response to questions beginning “How much do you want to know about …” 27 of 27 answered “Tell me all” to the questions about “your cancer,” “your prognosis,” “treatment benefits,” and “treatment side effects.” Only one of 27 answered otherwise: “Tell me a little” about cancer, and “Tell me some” about prognosis.

Secondary Outcomes

Participants were overoptimistic about the results of palliative chemotherapy, as shown in Table 3. Most (14/27, 52%) people thought a person with “metastatic cancer (breast, colorectal, lung, prostate—specific to that person's disease) spread to the bones and lymph glands” could be cured. After the decision aid, more people recognized that their cancer could not be cured (17/25, 63%) but eight of 25 (32%, P = 0.15, Fischer's exact test) still thought a person with metastatic disease could be cured. Patients were particularly overoptimistic about the chance of their symptoms being helped by chemotherapy: 87% thought their symptoms would be helped by chemotherapy, and 60% thought a patient would have at least 50% shrinkage of their cancer before the exercise, which declined only slightly after the decision aid. (While the correct answer varies by disease, the number helped by chemotherapy is usually less than 50%, and response rates are always less than 50%.)

There was no change in responses to the HHI after the intervention as we have previously reported.¹⁸ Participants did not appear to be visibly distressed by the intervention. A psychologist and chaplain were made available, but no one requested their services. In our small clinic, the primary nurses and doctors have frequent interactions during visits and chemotherapy. No patient was reported to be distressed in any way, during that visit or subsequent visits.

The comments recorded by the patients or the interviewers at the end of the exercise showed that most patients would share the information, as shown in Table 4.

In some cases, the average prognosis and treatment benefit, although small, was bigger than the person thought before the exercise. Nearly all found it helpful. Some illustrative comments are shown in Table 2.

We did not formally measure the time to complete the screening questions, pre- and posttests, pre- and post-HHI, and decision aid; but in most cases it took less than 20 minutes to complete the whole package including the pre- and post-tests. Review of the decision aid with the patient always took less than 5 minutes, even when we were reading it with the patient and family. This is consistent with work showing that oncologists state that completing an advance directive will take too much time but, in fact, it takes less than 10 minutes.^{[19] and [20]}

Discussion

Historical data show that patients know little about their prognosis and the effect that treatment will have on their cancer. Yet, this knowledge is essential to making informed choices about treatment benefits, risks, and even costs. When tested in randomized controlled trials, decision aids led to more involvement in decision making.^{[21] and [22]} However, there were no decision aids available about metastatic incurable disease, despite some promising early starts^{[23], [24], [25], [26], [27] and [28]} and only one about first-line treatment,²⁹ so we made a simple one. A successful decision aid may allow patients to discuss their situations with their physicians and develop management strategies that best concur with personal goals and preferences and help patients make plans in other areas of life.

Our findings suggest that most people do want honest information, even if the news is bad. We found that 27 of 27 enrolled patients initially reported wanting to know all the available information about their cancer, prognosis, treatment benefits, and treatment side effects. Also, 26 of 27 patients were able to complete the decision aid fully, our main outcome measure. While approximately 10% of available patients were excluded from accrual by their oncologists or oncology nurses due to preexisting distress, fear of distress in the patient or family member, uncontrolled symptoms, or psychiatric illness, in general there was excellent acceptance of the study by patients and oncologists. In this pilot study we did not investigate the attitudes of nonparticipants nor were we able to collect sociodemographic data to determine nonresponse bias, that is, whether certain types of patients are more likely to decline participation in the study.

Participants in the study were overoptimistic about their chances of cure, potential treatment response, symptom relief, and survival. None of these patients had curable disease, but 63% thought that a person with metastatic cancer of their type could be cured and gave the average chance of cure as 52%. Inaccurate assessment of cure rates decreased postintervention. At the pretest 14/27 (52%) believed a person with cancer similar to theirs could be cured, which changed to 8/26 (31%) at the posttest. This agrees with other studies that showed that patients mistook palliative radiation for curative radiation about one-third of the time, even when provided with accurate information.^{[1], [30] and [31]}

Knowledge of prognosis and planning for the future is important as there is evidence of benefit to having the discussion about treatment outcomes. Recent data show improved quality of care, improved quality of life, and improved caregiver quality of life if the physician discusses death with the patient and family.⁵ Transplant patients with advanced directives had more than a twofold survival advantage over those without them.²⁷ Conversely, over- or underestimating survival or treatment benefit can lead to bad health outcomes. Stem-cell transplant patients who were overoptimistic lived no longer than those with realistic views.^{[32], [33] and [34]} Cancer patients who overestimated their survival were more likely to die a “bad” death (defined as death in an intensive care unit, on a ventilator, or with multiple hospitalizations and emergency room visits) without achieving life extension.³⁵ It may be that the 16%–20% of patients with incurable solid tumors who start a new chemotherapy regimen within 2 weeks of death,³⁶ when they are unlikely to benefit, simply do not know the prognosis or treatment effect or have different perspectives.³⁷ Alternatively, we do not know how many patients decline second- and nth-line chemotherapy without knowing the full benefits and risks and who might choose chemotherapy if they knew second- or nth-line chemotherapy improved survival, pain scores, or quality of life. For instance, 40% of breast cancer patients will have some disease control from fourth-line chemotherapy for up to 4 months even if there is no evidence of improved survival.³⁸

Patients consistently tell us to be truthful, compassionate, and clear and to stay the course with them.^{[39] and [40]} Despite nearly all American patients stating that they want full disclosure about their prognosis, treatment options, and expected outcomes, most patients do not receive such information⁴¹ or receive such information far too late in their course.⁴² Even if terminally ill patients with cancer requested survival estimates, doctors would provide such estimates only 37% of the time, often an overestimate;⁷ and a recent meta-analysis showed that cancer physicians consistently overestimated prognosis by at least 30%.⁴³ Honest information respects the autonomy of a patient to make decisions based on what is known about the outcomes of such decisions.⁴⁴ Such information should not be forced on a patient, but the patient should be told that the information is available and that he or she has the right to accept or decline the information.⁴⁵

When we started this project, colleagues were concerned about whether patients would want such information, that patients would be distressed by poor prognosis, that patients would give up hope, and that the procedures would take too much time. We also were concerned about the effect of giving such bad news on the provider, when prior research showed negative effects on the information-giver's mood and affect from such encounters⁴⁶ and that doctors in general protect themselves by not giving bad news.⁴⁷ Completion of the decision aid was difficult for the interviewers, too. Some commented on how hard it was to give “bad” information about chance of cure and expected survival, even for patients they did not know. While patients may be more comfortable having advance directive discussions with a doctor they do not know rather than their oncologist,⁴⁸ it can still be hard for the provider. Surprisingly, it rarely took more than 20 minutes to discuss the information including the tests since the information was preprinted.

Patients vary in their approach to decision making, but the decisions should at least start with good information. Based on these preliminary findings, the piloted intervention is significant because it can lead to measurable impacts on knowledge about prognosis and appears to be judged helpful. We do not know the impact of full and truthful information on patient knowledge, decision making, hope, attendant choices about advanced medical directives, chemotherapy use, or hospice use. The next steps are to make the information available directly to patients on the Internet, which is in progress. The purpose is not to increase or decrease the use of palliative chemotherapy or hospice care; the lack of research into the decisions fully informed patients make precludes any such prediction. Since the intervention appears to be successful in this pilot trial, it will be tested in conjunction with standard care in a randomized clinical trial with measurement of quality of care, quality of life, and health-care cost outcomes.

Acknowledgments

This research was supported by VCU School of Medicine Research Year Out, GO8 LM0095259 from the National Library of Medicine (T. J. S., L. L., J. K.), and R01CA116227-01 (T. J. S.) from the National Cancer Institute.