PADM 5324

Speaking Notes

September 15, 2009

Dr. Neubauer

 

Yesterday, Mr. Patrick Swayze died of pancreatic cancer.  It had been about 20 months since diagnosis of his having the disease.  During that time he did what was possible to enhance the quality of his remaining life and to promote research efforts to find a cure for pancreatic cancer. 

 

Pancreatic cancer is the fourth leading cause of cancer death in the U.S. There are no early detection tools, few treatment options and no cure for this very difficult disease.

source: Julie Fleshman, JD, MBA
President and CEO
Pancreatic Cancer Action Network

http://www.pancan.org/announcements/swayze.html

 

Patients diagnosed with pancreatic cancer typically have a poor prognosis partly because the cancer usually causes no symptoms early on, leading to locally advanced or metastatic disease at time of diagnosis. Median survival from diagnosis is around 3 to 6 months; 5-year survival is less than 5%.[42] With 37,170 cases diagnosed in the United States in 2007, and 33,700 deaths, pancreatic cancer has one of the highest fatality rates of all cancers and is the fourth highest cancer killer in the United States among both men and women. Although it accounts for only 2.5% of new cases, pancreatic cancer is responsible for 6% of cancer deaths each year.[43]

source: http://en.wikipedia.org/wiki/Pancreatic_cancer

 

Chapter 6 -- Ways of Expressing Prognosis in Individuals (extended to populations)

 

People have attempted to identify their allotted times on earth in various ways for a very, very long time.  Some ancient methods strike us today as being rather odd.

http://en.wikipedia.org/wiki/Petosiris_to_Nechepso

 

There is a "natural progression" of a disease in an individual.

 

Point A -- biologic onset of the disease in an individual

Point P -- the evidence is there but no one has discovered the evidence

Point S -- development of symptoms and/or signs                                                                                              http://en.wikipedia.org/wiki/Sign_(medicine)

Point M -- the person (may) seek(s) medical care

Point D -- the patient receives a diagnosis (which hopefully is correct)

Point T -- the patient may receive some kind of treatment

Point O -- some outcome happens.  It may be cure, control, remission or death

 

If the outcome is death (as a result  of the particular disease identified) it is useful to accurately estimate PROGNOSIS. 

 

http://en.wikipedia.org/wiki/Prognosis

 

The use of the word, "prognosis" seems to presume death.  Most physicians and patients probably err on the side of optimism because of the possible psychological effect on the patient's attitude.

 

For policy purposes (and insurance and so forth) accuracy is probably more appropriate than optimism. 

 

            In order to evaluate new treatments it is important to have a meaningful (aggregate) "baseline" against which to measure.

 

THIS IS IMPORTANT.  The sooner new cases can be discovered the longer the apparent "survival time" will be even if no progress is made in terms of medical interventions.  As a result, SCREENINGS are likely to produce the appearance of longer survival times only because the condition was identified in patients earlier than would otherwise be the case.  This is called LEAD-TIME BIAS.  It is certainly true that EARLY DETECTION can improve prognosis, that that is a different aspect of measuring survival time.

 

Our textbook describes "many" measures of survival times of particular diseases.  I am going to refer to only five (5) of them.

 

1)         CASE FATALITY RATE http://en.wikipedia.org/wiki/Case_fatality_rate

 

Ideally suited to diseases that create short-term acute conditions leading to death.

Assumes that the (future) cause of death will be the disease/condition and not something else.

 

The explanation of case fatality rate at Wikipedia makes more sense to me than the explanation of it on page 110 of our textbook.

 

WITH NO TIMEFRAME it seems that this is the percentage of people who have a disease who don't eventually die of something else.  I don't see how this becomes an estimate of length of survival.

 

Given a time frame (from moment of diagnosis) it makes more sense to me. 

 

2)         FIVE YEAR SURVIVAL RATE

            http://en.wikipedia.org/wiki/Five-year_survival_rate

 

It is apparently the number of people still alive five years later divided by the number of patients in the cohort.  Those in the cohort were identified and began treatment in the same year.  For purposes of calculation it does not matter when in the five years some members of the cohort died.  Each person either survived for five years (and were counted in the numerator) or not.

 

3)         OBSERVED SUVIVAL RATE http://en.wikipedia.org/wiki/Survival_rate

 

It is not very interesting to follow only one cohort.  So you create a "Life Table."  EACH ROW is a different cohort of people who were newly treated in the same year.  For every row, the number still alive will tend to decrease as members of that cohort die of SOMETHING. 

 

A problem with survival rate studies is that patients tend to become "lost to follow-up" for reasons other than death.  "Dropping out" should not be treated as if they died, for purposes of calculating survival rates. 

 

3a)       AVERAGE SURVIVAL TIME

 

You have a cohort of people who were all identified and treated in the same initial year.  You follow them and record the date of their deaths.  When the last one of them dies, you calculate a average survival time.   Depending upon the disease, this may take a long time to calculate.  One or more people who survive an unusually long time may skew the average upward, depending upon the size of the cohort.

 

4)         MIDIAN SURVIVAL TIME http://en.wikipedia.org/wiki/AIDS#Prognosis

 

This is easier to calculate because you can make the calculation when half the members of a cohort have died.  It is less likely to be skewed upward by a few people who survive unusually long times.

 

5)         RELATIVELY SURVIVAL RATE http://en.wikipedia.org/wiki/Survival_rate

 

The explanation in our textbook gets pretty complex.  Basically, I think the idea is that survival rate should reflect the ages of the members of the cohort.  If the disease is not too immediately deadly, you would expect a cohort of younger people to on average live longer than a cohort of older people. 

 

Here is the explanation available at Wikipedia.

 

"Relative survival is calculated by dividing the overall survival after diagnosis of a disease by the survival as observed in a similar population that was not diagnosed with that disease. A similar population is composed of individuals with at least age and gender similar to those diagnosed with the disease."

 

CONCLUSION

 

Public Health efforts to estimate survival times associated with various diseases/conditions are based on aggregates of people.

 

This estimates can be useful in terms of public policy and planning.

 

The truth is that deciding a prognosis regarding an individual patient can be difficult because of the complexity of the human body and perhaps the results of attitude.

 

In the aggregate, studies based upon large numbers of people are preferable to studies based on small numbers of people.  Exceptional cases are less likely to skew the findings of a large study than a small study.