Download Logistic Regression Analyses: The 13 measures related to the

-1Supplemental Materials for Why Older Adults Make More Immediate
Treatment Decisions About Cancer Than Younger Adults
The Supplemental Materials include a series of logistic regressions analyses for Study 1
and Study 2. These are followed by situations 1 and 2 of the prostate cancer scenario used in
Study 1 and situation 2 of the breast cancer scenario used in Study 2.
Logistic Regression Analyses
A series of logistic regression analyses were conducted on the data from both studies and
provided good confirmation of the decision tree results. The analyses allowed for investigation of
the effects of numerous knowledge, interest, and resource variables as well as their combined
ability to account for age effects in the tendency to make immediate treatment decisions. For
each study the best knowledge, interest, and resource predictor variables were stepped into a
final logistic regression equation in the order predicted by the model for decision making about
cancer treatments displayed in Figure 1 of the Psychology and Aging article.
Logistic Regression Analyses for Study 1 about Prostate Cancer
Logistic regression was used to examine the 13 measures previously examined with
decision tree analysis and related to the decision-maker variables of knowledge, resources, or
interests in our model (see Figure 1). The same three measures of knowledge about cancer or
prostate cancer entered as predictors in the decision tree analysis were entered in a logistic
regression analysis to predict immediate versus delayed treatment decisions. Treatment
knowledge solely for prostate cancer (Wald = 4.56, df = 1, p < .033) contributed significant
predictive power, while two other knowledge measures (total knowledge for prostate cancer;
total knowledge for diagnosis and treatment for prostate cancer and other types of cancer) did not
contribute significantly. The more prostate cancer treatment knowledge possessed by a
-2participant, the more likely an immediate decision about a treatment was made when presented
the diagnosis and treatment options by the physician in the scenario. The statistical model
correctly predicted 89.9% of immediate decisions and 25.5% of delayed decisions; Nagelkerke’s
R2= .06 (Nagelkerke, 1991). Both the decision tree analysis and the logistic regression pointed to
greater knowledge of prostate cancer treatments leading to immediate treatment decisions,
lending support to our model (see Figure 1).
A second logistic regression looked at the six interest predictors and found that interest in
the scenario prior to situation 2 (Wald = 6.78, df = 1, p = .009), but not the more trait-like interest
measures of the MBSS (Miller, 1987) or the PSDM (Deber et al, 1996), predicted timing of the
decision with higher interest exhibited in the scenario related to delaying the decision. The
statistical model with interest correctly predicted 81.2% of immediate decisions and 40.4% of
delayed decisions (Nagelkerke’s [1991] R2= .14).
The third logistic regression looked at the three cognitive resources predictors of
education, working memory and vocabulary. Vocabulary (Wald = 3.66, df = 1, p = .056) added
some predictive power, while education (Wald = .12, df = 1, p = .72) and working memory
(Wald = .46, df = 1, p = .50) did not approach significance. The statistical model with these
predictors correctly predicted 81.2% of immediate decisions and 25.5% of delayed decisions
(Nagelkerke’s [1991] R2= .06). The higher the vocabulary the more likely a delayed treatment
decision was made; this is the same relationship observed at nodes 3 and 4 of the AnswerTree
cluster analysis seen in Figure 2. A logistic regression with only age group as a predictor showed
age group to be a significant predictor of immediate versus delayed treatment decisions (Wald =
4.61, df = 1, p = .032). The statistical model with age group correctly predicted 82.6% of
immediate decisions and 23.4% of delayed decisions (Nagelkerke’s [1991] R2= .06).
-3Finally, the significant predictors from the above logistic regression analyses plus other
predictors identified in the decision tree analysis were entered in one analysis and stepped into
the equation in the order predicted by our model (knowledge, interest, cognitive resources) and
followed by age to see if they could explain differences in decision time attributed to age. The
findings are displayed in Table 1. There was a trend for more prior treatment knowledge relating
to immediate decisions about prostate cancer treatment. The knowledge predictors shown in
Table 1 yielded a model with Nagelkerke’s (1991) R2= .04; knowledge could predict about 4%
of the variance and not as much as anticipated by the model in Figure 1.
The state measure of interest was the only significant interest variable. As shown in
Table 1, the addition of the interest variables to the knowledge predictors improved the
prediction of the model so that it could predict about 11% more variance (Nagelkerke’s [1991]
R2= .15). The addition of the cognitive resource variable of vocabulary also improved the
prediction of the model (Nagelkerke’s R2= .20). The direction of relationships of these
knowledge, interest, and cognitive resource variables are compatible with our model (see Figure
1). As seen in Table 1 these variables could account for all of the age effects. In the final logistic
regression model 21% of the variance in immediacy of treatment decision-making could be
explained (Nagelkerke’s R2= .21).
Logistic Regression Analyses for Study 2 About Breast Cancer
Logistic regression was used to examine the 19 measures initially examined with decision
tree analysis and related to the decision-maker variables of knowledge, resources, or interests in
our model (see Figure 1). The same seven measures of knowledge about cancer or breast cancer
entered as predictors in the decision tree analysis were entered in a logistic regression analysis to
predict immediate versus delayed treatment decisions. Prior knowledge of breast cancer
-4treatments (Wald = 19.16, df = 1, p < .0005) and misconceptions about breast cancer (Wald =
7.80, df = 1, p = .005) as measured by two subtests of the Vaeth’s (1993) test for knowledge of
breast cancer and self-reported ratings of knowledge of cancer (Wald = 6.67, df = 1, p = .010)
added significant predictive power, while self-reported knowledge of breast cancer and the other
scores from the Vaeth test did not. The more treatment knowledge possessed by a participant and
the fewer the misconceptions about breast cancer, the more likely an immediate decision about a
treatment was made when presented the diagnosis and treatment options by the surgeon in the
scenario. The statistical model with these significant predictors correctly predicted 68.2% of
immediate decisions and 67.1% of delayed decisions; Nagelkerke’s R2= .22 (Nagelkerke, 1991).
Both the decision tree analysis and the logistic regression pointed to greater prior knowledge
about breast cancer treatments leading to immediate treatment decisions, lending support to our
model (see Figure 1).
A second logistic regression looked at the interest predictors and found that self-reported
ratings of interest for cancer (Wald = 4.263, df = 1, p = .039), but not for breast cancer (Wald =
.43, df = 1, p = .51) predicted decision time with higher interest in cancer related to delaying the
decision. The direction of the relationship is the same as that noted for high knowledge, older
women in the cluster analysis. The statistical model with interest in cancer as a predictor
correctly predicted 47.2% of immediate decisions and 64.8% of delayed decisions (Nagelkerke’s
[1991] R2= .03).
The third logistic regression looked at the nine cognitive resources predictors.
Vocabulary (Wald = 5.04, df = 1, p = .025) and forward digit span (Wald = 4.63, df = 1, p = .03)
added significant predictive power, while reading skills (Wald = 3.31, df = 1, p = .069),
education (Wald = .04, df = 1, p = .84), reaction time (Wald = .49, df = 1, p = .48), MMSE (Wald
-5= .73, df = 1, p = .39), digits backward (Wald = .30, df = 1, p = .59), RSPAN (Wald = .75, df = 1,
p = .39), and CSPAN (Wald = 1.68, df = 1, p = .20) did not. The statistical model with these
predictors correctly predicted 66.7% of immediate decisions and 66.3% of delayed decisions
(Nagelkerke’s [1991] R2= .15). The higher the vocabulary the more likely an immediate
treatment decision was made; the older adults scored higher on the vocabulary test than the
younger, but the variability in the performances among the older group was also larger (Old: M =
56.23, SD = 21.17; Young: M = 43.17, SD = 11.66). The more diminished adults’ memory
performance (i.e., digits forward) and lower their reading comprehension skills (i.e., Davis
Reading Test), the more likely immediate decisions were made. Meyer et al. (1995) found that
types of treatments selected were related to prose processing; poor memory and reading skills
can handicap treatment decision-making. Both the decision tree and the logistic regression
analyses pointed to cognitive resources variables impacting decision time, but the specific
variables differed. Due to the linear constraint of logistic regression, the education variable
would have been missed without the earlier subgroup analyses since its relationship to decision
time appears to be curvilinear.
A logistic regression with only age group as a predictor showed age group to be a
significant predictor of immediate versus delayed treatment decisions (Wald = 18.57, df = 1, p <
.0005). The statistical model with age group correctly predicted 64.6% of immediate decisions
and 67.4% of delayed decisions (Nagelkerke’s [1991] R2= .13).
Finally, the significant predictors from the above logistic regression analyses plus other
linear predictors identified in the decision tree analysis were entered in one analysis and stepped
into the equation in the order predicted by our model (knowledge, interest, cognitive resources)
and followed by age group to see if they could explain differences in decision time attributed to
-6age. The findings are displayed in Table 2. The more treatment knowledge possessed by
participants, the more likely they made an immediate decision about breast cancer treatment.
However, the more general cancer knowledge they possessed as measured by self-reported
cancer knowledge or total scores overall about breast cancer, the more likely they were to delay
their treatment decision. Perhaps participants’ general knowledge pointed out the seriousness of
a cancer diagnosis and the need to delay their decision to find out more about this particular type
of cancer and its various options for treatment. The significant knowledge predictors shown in
Table 2 yielded a model with Nagelkerke’s (1991) R2= .20.
Both self-rated knowledge about cancer and interest in cancer, rather than specific
knowledge about breast cancer, were related to waiting to make a treatment decision. This
suggests that participants with interest or knowledge about cancer in general preferred to wait to
make a decision so that they could gather information related to this specific type of cancer,
breast cancer. As shown in Table 2, the addition of the interest variable to the three knowledge
predictors slightly improved the prediction of the model (Nagelkerke’s [1991] R2= .22).
The addition of the cognitive resource variables also improved the prediction of the
model (Nagelkerke’s [1991] R2= .25). As can be noted in Table 2, only CSPAN was a significant
predictor. The higher the working memory of the participant, the more likely they were to delay
their treatment decision. The direction and magnitude of relationships of these knowledge,
interest, and cognitive resource variables are compatible with our model (see Figure 1). As seen
in Table 2, these variables could account for most but not all of the age group effects; age group
was still a significant predictor of timing for treatment decisions and its addition to the model
increased the Nagelkerke’s R2 from .25 to .28.
-7Table 1
Summary of Logistic Regression with Predictor Variables Stepped into the Equation in the Order
of Prior Knowledgea, Interestb, and Cognitive Resourcesc Variables
Variables
B
SE B
Wald
df
p
-.14
.08
2.89
1
.09
.11
.67
2.55
1
.11
-.09
.09
1.61
1
.28
All cancer D & Tx
.07
.07
.88
1
.35
Blunter
.06
.10
.31
1
.58
Scenario interest
.18
.08
4.98
1
.03
PSDM problem
.58
.41
2.05
1
.15
PSDM decision
.22
27
.64
1
.43
-.12
.09
1.67
1
.20
All cancer D & Tx
.05
.07
.50
1
.48
Blunter
.04
.10
.18
1
.67
Scenario Interest
.17
.08
4.72
1
.03
PSDM problem
.62
.42
2.20
1
.14
PSDM decision
.15
.28
.284
1
.60
Vocabulary
.03
.02
4.09
1
.04
Step 1– Knowledge
PCa Tx
All cancer D & Tx
Step 2 – Interest
PCa Tx
Step 3 – Resources
PCa Tx
Step 4 – Age*
-8PCa Tx
-.10
.09
1.09
1
.30
All cancer D & Tx
.05
.07
.50
1
.48
Blunter
.02
.11
.03
1
.87
Scenario interest
.16
.08
3.74
1
.05
PSDM problem
.57
.42
1.87
1
.17
PSDM decision
.06
.29
.04
1
.84
Vocabulary
.03
.02
4.46
1
.04
-.01
.01
.89
1
.35
-.3.55
1.79
3.96
1
.05
Age
Constant
*Nagelkerke’s (1991) R2= .21; Correctly predicts 82.6% immediate decisions and 48.9% delayed
decisions.
a
PCa Tx = treatment knowledge of prostate cancer; All cancer D & Tx = diagnosis and treatment
knowledge of prostate cancer and other cancers.
b
Blunter = score from the blunting subscale of the Monitor/Blunter Style Scale (Miller, 1987);
Scenario interest = amount of interest shown via search depth in the opening situation of the
prostate cancer scenario; PSDM problem = score from the problem-solving subscale of the
Problem-Solving Decision-Making Scale (Deber et al., 1996) and PSDM decision = score from
the decision-making subscale.
c
Vocabulary = score from the Quick Word Test (Borgatta & Corsini, 1964).
-9Table 2
Summary of Logistic Regression with Predictor Variables Stepped into the Equation in the Order
of Prior Knowledgea, Interestb, and Cognitive Resources Variablesc
Variables
B
SE B
Wald
df
p
.45
.10
16.81
1 .000
Breast cancer total
-.07
.03
6.50
1 .011
Ratings for cancer
-.38
.16
5.84
1 .016
Misconceptions
Out
Step 1– Knowledge
Treatments
.274
Step 2 – Interest
Treatments
.46
.11
18.00
1 .000
Breast cancer total
-.07
.03
6.45
1 .011
Ratings for cancer
-.32
.16
4.16
1 .042
Interest in cancer
-.21
.13
2.80
1
.47
.11
18.58
1 .000
Breast cancer total
-.07
.03
6.51
1 .011
Ratings for cancer
-.36
.16
4.95
1 .026
Interest in cancer
-.21
.13
2.86
1 .091
CSPAN
-.27
.13
4.51
1 .034
Digits forward
Out
.497
Digits back
Out
.715
094
Step 3 – Resources
Treatments
- 10 Vocabulary
Out
.130
Reading skill
Out
.889
Step 4 – Age*
Treatments
.42
.11
13.61
1 .000
Breast cancer total
-.06
.03
4.99
1 .025
Ratings for cancer
-.37
.16
5.14
1 .023
Interest in cancer
-.21
.13
2.71
1 .100
CSPAN
-.17
.14
1.51
1 .219
.77
.39
3.82
1 .051
1.54
1.33
1.34
1 .248
Age group
Constant
*Nagelkerke’s (1991) R2= .28; Correctly predicts 72% immediate & delayed decisions
a
Treatments = Vaeth’s Breast Cancer Treatment Knowledge (Vaeth, 1993); Misconceptions =
Vaeth’s Breast Cancer Misconception Knowledge; Breast cancer total = Vaeth’s Breast Cancer
Knowledge Total Number Correct; Ratings for cancer = Self-rated knowledge of cancer
b
c
Interest in cancer = Self-rated interest in cancer
Digits forward and Digits backward from WAIS-R (Wechsler, 1981); CSPAN = Computational
Span (Babcock & Salthouse, 1990); Vocabulary = score from the Quick Word Test (Borgatta &
Corsini, 1964); Reading skill = score derived from the Davis Reading Test
(Davis, 1944)
- 11 Prostrate Cancer Scenario Used in Study 1: Situations 1 & 21
Situation 1 – The Decision Task Format
The scenario is set within a hypertext environment so that you may access new information any
time you see blue or purple underlined hyperlink. One of the beauties of hyperlinks is that you
do not need to double click on them. Simply click once with the left mouse button and you will
be taken to the corresponding page. If you are accustomed to "surfing," we ask that you not use
the browser's navigation bar located at the top of the page. Instead, please use the navigation
tools (hyperlinks) provided for you within each page. All pages have links at the bottom that you
are instructed to click when you are finished with the information provided on the page.
Following these links will enable you to access other information or other parts of the decision
scenario. If you get confused about where you are in the scenario or in the web pages, you may
use the back button at the top of the page. However, this should only be done if you can see no
other means of finding where you are.
Ready To Begin: Your New Identity
Now that you are ready to begin, we would like to ask you to assume a new identity. Please
imagine that you are an active, married man. You just turned 60 and are very healthy. Try to
keep this in mind as you make this medical decision. However, remember that even though your
age may have changed, we want to know how you make decisions, so be true to your way of
doing things.
Just so that we know you are ready to begin, turn to the experimenter and explain to him
what you are supposed to do for this decision task.
- 12 Begin
News From Your Latest Physical
During your last routine physical exam your doctor noticed an unusual growth in your prostate.
You were referred to a urologist after your initial PSA (prostate specific antigen) test came back
higher than normal for a man your age. The urologist performed a series of tests (including
another digital rectal exam (DRE), PSA tests and a trans-rectal ultrasound (TRUS)) to confirm
the original results. All of the tests confirmed that something was abnormal with your prostate.
What Is It?
The urologist explained that there are a couple of conditions that can produce these results. The
most common is called benign prostatic hyperplasia (BPH). This is a non-cancerous growth of
the prostate. He explained that this growth is a normal part of the aging process and that when
the prostate grows faster than expected it can produce elevated PSA levels. The other condition
that could be causing the results is a cancerous growth within the prostate. Prostatic carcinoma
is the second leading type of cancer among men (skin cancer is the most common). In addition,
it is the second leading cause of cancer deaths among men (behind lung cancer).
The urologist explained that the TRUS results indicate that the PSA level is really higher
than would be expected given the size of your prostate (high PSA density), and that it appears as
though a small cluster of cells are responsible for the elevated PSA. He said "Now it could still
be a benign growth in your prostate that is localized in this area. However, I need to tell you that
it is more probable that we are talking about a malignant growth. The reason I say this is
because of the results of the free-PSA test.
- 13 "This test looks for the proportion of PSA floating in your bloodstream unaccompanied
by other blood proteins. The accepted cutoff for normal levels of free PSA is 25%. Less than
that usually signifies cancer. Your percentage free-PSA is right around 22. By itself, this count
isn't that far off normal, but in conjunction with the abnormal DRE, the elevated PSA, and the
high PSA density it looks like the possibility of cancer is pretty high."
He insisted that even with all of the test results they have, it is impossible to know what
your condition is without a biopsy. He explained that the only way to make a firm diagnosis of
your condition was to have a pathologist analyze cells from the growth and the surrounding
prostate tissue. He performs biopsies in his office. The procedure is very short and can be done
with or without a local anesthetic. He would like to set up the biopsy as early as schedules will
allow.
What would you do?
1) Schedule the biopsy
2) Delay the scheduling of the biopsy
(Next Page for those who scheduled biopsy)
- 14 You have scheduled your biopsy for next week. What would you like to do until then?
1) Nothing, try not to think about it
2) Seek information about biopsies. (Information available: newspaper, 2 Internet sites, 1 video)
3) Talk to your spouse
4) Talk to other family members and close friends.
5) Seek spiritual counsel (from clergy or others)
Situation 2 – Biopsy Results
Your biopsy went fairly well. You were amazed at how much pain could be generated by
a single needle (you opted for the localized anesthesia and, boy, are you ever thankful you did!).
You waited four days for the doctor's office to call with the results of the biopsy. The news was
not what you wanted to hear, prostatic carcinoma. The urologist scheduled a consultation
appointment.
At the consultation the urologist explained that the growth was indeed malignant. He
explained that cancer cells are rated in a couple of different ways. The most commonly referred
to rating scale is the one used to indicate the stage of the cancer. Stage I indicates that the cancer
cells are confined to the prostate and that they are too small to be detected by any means other
than a biopsy. Stage II is cancer that is still confined to the prostate, but is large enough to be
felt during a DRE. Cancer at either of these two stages is called "early stage cancer." These
cancers are the most treatable because the cancer has not progressed outside of the host site.
- 15 Stage III indicates cancer that has invaded surrounding areas. Stage IV cancer is the
latest stage of cancer. It entails the movement of cancer cells out of the area and into the
bloodstream. At this stage cancer is referred to as having metastasized. Once the cancer has
progressed to this stage it grows in other regions of the body.
The urologist explained that there is no evidence that the cancer has spread outside of the
prostate. While the tumor is small, it is considered stage II because it was found through the
DRE. He further explained that the other commonly used way of describing cancer cells is
through "grading" using a Gleason score. The Gleason score describes how different the cells
are from normal cells. The scale goes from 2 to 10. More abnormal cells receive higher grades.
Your biopsy was graded as a 4. This is still considered low.
He explained that this was a great sign, because it meant that the cancer was not
aggressive. However, there is no way of knowing how long it will take to progress into a more
dangerous form of the disease. At this point, the doctor explained the options you have for
treatment.
Treatment Options
The urologist explained that there are three main types of treatment for early stage prostate
cancer. These include radical prostatectomy, radiation therapy, and expectant therapy (watchful
waiting). He explained that all three options had arguments for and against them. Each one had
the potential for side effects and none could guarantee the elimination of the cancer.
Radical Prostatectomy
Radical prostatectomy is the most invasive means of treatment. He explained that this treatment
involves a one and a half to three hour surgery. During surgery he would first take a biopsy of
the lymph nodes in the pelvic region. Assuming the pathology report says the lymph nodes are
- 16 cancer free, he will take out the prostate gland while trying to preserve as much of the soft tissue
and nerves surrounding the prostate as possible. Since the tumor is small, the chances of
eliminating the cancer are very good. Since the treatment requires a major surgery, there is a
slight chance (.5% to 3%) of death. However, the main side effects for this treatment are
incontinence and impotence. While most men experience short-term problems with both, the
long-term impotence occurs in 30% to 90% of patients. The large range mainly depends on how
much of the nerve is spared and how much pretreatment difficulty a patient has in achieving an
erection. Incontinence is universal for a short period after the catheter is removed. However,
only 32% of men report any further incontinence, and only 7% of men report complete
incontinence.
Radiation Therapy
Radiation therapy, on the other hand, has a lower risk of death with comparable or greater risk of
side effects such as impotence and incontinence. Specifically, he explained, there is a 0.2%
chance of death from the treatment. The rate of impotence is around 40%. Sixty percent of men
report some degree of incontinence, while only 1% report complete control loss.
Of the other potential side effects, blockage of the urethra (called a urinary stricture) and
rectal injury are the most common. Both treatment options involve risk of these effects. Surgery
causes urinary stricture in 12% to 20% of patients, and causes rectal injury in approximately
30%. However, he noted, radical prostatectomy done with the retropubic (coming in from the
abdomen region) method entails lower risk of rectal injury. Radiation therapy results in urinary
stricture in about 5% of patients. Rectal injury occurs in approximately 11% of patients.
The doctor hands you a few pieces of literature which detail the numbers he explained to
you. One piece, Understanding Treatment Choices for Prostate Cancer, also contains
- 17 information about other aspects of the treatments. He then explained that the research is still
unclear about which treatment is best for your type of prostate cancer. However, he cited
research that suggests that if you are going to live longer than 10 to 15 years, the probability of
reoccurrence is lower when a radical prostatectomy is chosen. He also reaffirmed his belief that
because of the small nature of the tumor, there was a good likelihood that he would be able to
spare half of the nerves that allow erections (the half located on the side of the prostate that is
tumor free).
Watchful Waiting
He then explained that there was a third option, watchful waiting. This is essentially a gamble
that the cancer will progress slowly. If this option is chosen, the object will be to monitor the
PSA levels so that if there is any indication that the tumor has grown, an alternate treatment can
be selected. Of course, there are two big problems with this approach.
First, there is no way of knowing how long it will take for the cancer to grow or
progress. The cancer might be very slow growing and might allow for decades of normal living
prior to causing any difficulty. However, it might spread so fast that the window of treat-ability
could be missed. In essence, he explained, this is a gamble. On the other hand, you have the
benefits of avoiding all of the potential side effects that are caused by the other treatments. That
is, you avoid them until you need to choose an alternative treatment option.
The main reason that watchful waiting is a plausible treatment option is that there is a
lack of research findings indicating an increase in survival rate between treatment and nontreatment of prostate cancer. The reality is that the decision is a difficult one. You need to
weigh the potential risks and benefits of each option and decide what way you want to treat this
disease.
- 18 Would you like to choose an option, now?
Yes, I am ready to choose an option.
No, I need more time
1From How much information do men really want? Information search behavior and decision
rationale in a medical decision-making task for men (pp. 110-113; 116-118) by A. P. Talbot,
2004, Unpublished doctoral dissertation, The Pennsylvania State University, University Park.
Copyright 2004 by A. P. Talbot. Reprinted with permission.
Breast Cancer Scenario Used in Study 2: Situation 22
You opted for the lumpectomy immediately, and the lump was not cancerous. However,
when on the operating table the surgeon had probed the area with her fingers and felt a smaller
lump that had not appeared on the mammogram or the ultrasound. This other 8-mm (1/3 in.)
lump was removed. This small lump was intraductal carcinoma insitu; that is, with the light
microscope and the cells sampled there was no evidence of the cancer moving out of the
mammary ducts and into surrounding breast tissue. (Intraductal carcinoma means that the cancer
is in the milk ducts and in situ means that the pathologist did not see evidence for the spread of
the cancer outside the duct membrane.) On one side of this cancerous lump within the duct,
slightly less than 1mm (0.04 in.) of surrounding noncancerous tissue was available for analysis,
perhaps this side of the lump faced the other lump of tissue removed, but perhaps this side faced
the tissue still in your breast. However, in the sampled cells from the 1mm available to
examine no cancer cells were seen outside of the cancerous duct.
- 19 The surgeon tells you that she does not know what she would do in your situation. She
thinks that she would either have a mastectomy (breast removal) or do nothing else
but have frequent mammograms and checkups. The options she gave you are listed on the next
page. She explains that she will be leaving town tomorrow for 2 weeks andthat you will need to
make your own decision. She will be happy to have her secretarial staff set up any appointments
with cancer specialists, and so on, if you are interested in speaking to them. She explains that
experts disagree about what to do with the kind of cancer you have. She explains that they used
to do mastectomies routinely for the spreading (invasive) type of cancer, but recent research has
shown equal success rates for lumpectomy and radiation. With the in situ cancer, they used to
routinely do mastectomies too, but there hasn't been much research on in situ cancer because it is
less common. The experts do not have an answer for your health problem, and you will
have to find your own personal answer. Here are the options she gave you:
1. Removal of both breasts.
2. Removal of the left breast.
3. Having another lumpectomy (reincision) to see if there is any more cancer in the area where
the small cancerous lump was found.
4. Radiation to sterilize the area where the lump was found.
5. Random assignment to a national study of intraductal carcinoma in situ: half of the people are
assigned to radiation and half are assigned to no treatment after both groups
had the initial lumpectomy when the cancer was found.
6. Nothing but 6-month mammograms and checkups for the rest of your life.
Decision: Would you make a decision at this point? ______________
Why? ________________________________________________________________________
- 20 _____________________________________________________________________________
If not, what would you do next? ___________________________________________________
_____________________________________________________________________________
What kinds of information would you try to get to help you make this decision? _____________
_____________________________________________________________________________
Write your plan for what you would do next? _________________________________________
_____________________________________________________________________________
2From “Discourse comprehension and problem solving: Decision about the treatment of breast
cancer by women across the life span.” By B. J. F. Meyer, C. Russo, and A. Talbot, 1995,
Psychology and Aging, 10, p. 101. Copyright 1995 by the American Psychological Association.
Reprinted with permission.