Prostate cancer is the second leading cause of cancer related death among men in the United States, and accounts for 29% of all cancers diagnosed in men. Furthermore, approximately one in six men will be diagnosed with prostate cancer in their lifetime. Thus, 17% of male Veterans will be asked to make a decision about the treatment of their prostate cancer. The burden of this disease is further magnified when one considers that most patients will live for years following their diagnosis and with any adverse effects of therapy. Given that there have been no clinical trials showing that any prostate cancer treatment produces an increased likelihood of survival; men are asked to actively participate in treatment decisions. Previous research has revealed that men are often uninformed about their prostate cancer, particularly African American men and men with lower educational attainment. Thus, it is critical to develop and test decision aids that can help all men (especially men with low literacy skills) make an informed decision.
The goal of the study was to compare the impact of a plain language decision aid (DA) to a conventional DA on prostate cancer patients' decision making experience and communication with their physician.
Specific Aim 1: To test whether the plain language DA increased the quality of the decision making process beyond that provided by a conventional decision aid. We hypothesized that the plain language DA would increase men's knowledge, interest in shared decision making, and improve men's perception of patient-physician communication.
Specific Aim 2: To evaluate the impact of the plain language DA on patient-physician communication and trust. We hypothesized that the plain language DA would increase trust, as well as discussion of treatments (especially side effects) and shared decision making.
Specific Aim 3: To test whether the plain language DA is effective in both low and high literacy patients. We hypothesized that the plain language DA would be effective in both high and low literacy patients.
Specific Aim 4: To test whether the plain language DA is effective in both African American and White patients. We hypothesized that the plain language DA would be effective in African American and White men.
1028 men without a prior history of prostate cancer and undergoing a prostate biopsy were recruited from 4 VA hospitals and randomized to receive one of two decision aid booklets (plain language vs. conventional). Additional inclusion criteria included ability to speak English, provide informed consent, and have a PSA < 20. At Time 1, which occurred prior to diagnosis, all participants completed a baseline interview (N=1023). Men diagnosed with localized prostate cancer completed two additional interviews (N=334). The Time 2 interview was conducted prior to the patient receiving his cancer diagnosis from his physician (N=285). The treatment discussion between patients and their physician was audio recorded (N=258). The Time 3 interview was conducted over the phone 7-10 days following the diagnosis visit (N=244). The surveys included measures of literacy, numeracy, anxiety, preference for shared decision making, knowledge, treatment preferences, risk perceptions, perception of patient-physician communication, and confidence and satisfaction with the decision making process. All survey questions were read aloud and responses recorded. One-way analyses of variance (ANOVA) and Chi-squared tests were conducted to compare the effect of decision aid type on outcome variables.
Participants receiving the plain language DA demonstrated greater prostate cancer knowledge (F=5.43, p = 0.02). Among those diagnosed with cancer, there were no differences in interest in shared decision making by DA at Time 1 (3.21 vs. 3.29, F=0.93, p>0.20) or Time 3 (3.63 vs. 3.52, F=2.89, p>0.20). However, after receiving a DA (but before receiving their cancer diagnosis), those receiving the plain language DA were more interested in having an active role in the decision than those who received the conventional DA (3.50 vs. 3.33, F=5.10, p=0.025).
There were marginal differences in patients' treatment preferences by DA at Time 2 ( 2=12.32, p=0.055) and Time 3 ( 2=11.18, p=0.08). The largest difference was in interest in active surveillance at Time 2. Whereas 39.8% of those receiving the plain language DA considered active surveillance, only 20.7% of those receiving the conventional DA considered it.
There were no differences in confidence and satisfaction with the decision making process (all p's>0.20), trust (p>.20), or anxiety (all p's>0.25) by DA type.
There were no differences by race in terms of knowledge, satisfaction/use of DAs, preference for shared decision making, perception of patient-physician communication, or trust in the VA. However, African Americans expressed more anxiety at Time 1 (1.09 vs. 0.95, F=5.25, p<0.01) and less interest in active surveillance pre-diagnosis (22.8% vs. 42.2%, p<0.02).
Lower literacy individuals were more anxious at all time periods (p<0.01), less interested in shared decision making before receiving a DA (F.25 vs. 2.38, F=8.08, p<0.01), and in active surveillance (23.8% vs. 41.8%, p<0.02).
In 97% of appointments, urologists discussed surgery and radiation as treatment options compared to 89% where active surveillance was discussed. The risks of incontinence and impotence with surgery were discussed with 88% of patients, whereas the risk of incontinence with radiation was discussed with 73% of patients, and 56% received information about impotence with radiation. 47% of patients were told about the impact of radiation on bowel function. Whether these topics were discussed did not differ based on patient race or literacy. However, for patients who indicated prior to the appointment that they wanted to be the primary decision maker (rather than sharing the decision with their physician or having their doctor make it), physicians were less likely to discuss surgery (p<0.05), radiation (p<0.05), and active surveillance (p<0.08).
We were underpowered to examine interactions between the type of DA and race or literacy.
This study adds significant value to the field. Most studies of DAs are only conducted at one site and are limited in the generalizability of the data. In contrast, this study was conducted in 4 regions of the United States (East, South, Midwest, and West). Additionally, very few studies have tested the impact of a DA on patient-physician communication. This study recorded the doctor telling a patient his cancer diagnosis, which will help illuminate how DAs affect actual communication. By measuring the literacy of participants we were able to determine the impact of literacy on our outcome variables. Finally, we were mindful of rates of African Americans at our different sites and conducted analyses to determine if there were ethnic or racial differences in our outcome variables.
- Levy MH, Back A, Benedetti C, Billings JA, Block S, Boston B, Bruera E, Dy S, Eberle C, Foley KM, Karver SB, Knight SJ, Misra S, Ritchie CS, Spiegel D, Sutton L, Urba S, Von Roenn JH, Weinstein SM. NCCN clinical practice guidelines in oncology: palliative care. Journal of the National Comprehensive Cancer Network. 2009 Apr 1; 7(4):436-73.
- Fagerlin A, Holmes-Rovner M, Knight S, Ling B, Alexander S, Tulsky J, Rovner D, Tobi JE, Kahn VC, Ubel PA. Literacy and numeracy in Veterans and their impact on cancer treatment perceptions and anxiety. Poster session presented at: Society for Medical Decision Making Annual Meeting; 2011 Oct 22; Chicago, IL.
- Scherer LD, Ubel P, Holmes-Rovner M, Knight S, Alexander S, Ling B, Tulsky J, Abir M, Fagerlin A. Literacy and irrational decisions: Bias from beliefs, not from comprehension. Paper presented at: Society for Medical Decision Making Annual Meeting; 2011 Oct 22; Chicago, IL.