In March 1998, I suggested to Barbara Lackritz,(alias GrannyBarb) list manager of the CLL List, that a survey of list subscribers with Chronic Lymphocytic Leukemia (CLL) would be interesting to me and, I thought, to other subscribers. I had nothing fancy in mind but rather an effort to learn some of the facts about us collectively, like our ages and where we live, how long we've had CLL, what treatments we've undertaken. After consulting with several others, she agreed, and during April, solicited and organized potential questions. Not unexpectedly, more questions were proposed than most people could be expected to answer at any one time. Other questions proved difficult to state in a form that could be answered through a brief, written questionnaire. Still others seemed to deal with matters better explored through a survey of patients' records than through patients' memories.

After some email exchanges, we agreed on a draft questionnaire, and in late April GrannyBarb sent it to 20 volunteer "beta testers." Their answers and comments led to revisions. In late May, the final version of the questionnaire was distributed by email to list subscribers, and answers began to come in. The questionnaire was distributed at least twice more, and in mid-June, a web version of the questionnaire was also posted. Data collection stopped July 10, although a couple of questionnaires received after that date have been included.

In the roughly six weeks that the questionnaire was available, 340 responses were received with useable information about someone with CLL--275 from CLLers themselves, and 65 from a spouse, son or daughter, friend, sibling, parent or grandchild. The information about these CLLers (whom we shall sometimes call "respondents," disregarding the fact that some responses were sent by helpers) provides the basis for the reports to follow. Each report will be quite brief, as befits our attention span in this publication medium. Later, the reports will be collected and integrated into one overall document.

Since the reports will be coming from me, it is appropriate that I introduce myself. I am Sheldon Messinger, a CLLer, diagnosed in December 1997 at Stage Zero and, at this writing at least, still without treatment. My education is in sociology. I did research and taught at the University of California, Berkeley most of my working life. (I retired in 1991, although I still work at the university part-time.) I've had some training in, and some experience with, designing and interpreting surveys, but I am not an expert in their construction or analysis. I have sought and received much assistance with this survey.

As I've already mentioned, GrannyBarb consulted with several persons to develop questions. Michael Pearlman, a fellow CLLer on this list, developed the web version of the questionnaire, which some of you used. Iona Mara-Drita, a Ph.D. candidate in sociology at Berkeley, helped process the data. Three friends and colleagues have provided help in diverse ways. John Berecochea taught me most of what I know about coding data; he helped develop the codes used in the CLL survey. S. Leonard Syme, an expert in social epidemiology, continues to provide advice about the what analyses make the most sense given the character of the survey data. Richard A. Berk, an expert in social statistics, has provided guidance every step of the way on what statistical procedures to use and how, cautiously, to interpret the results. (Berecochea is currently Director of Research, California Department of Corrections; Syme recently retired from UC, Berkeley; Berk continues to teach at UCLA.) Each of these persons has contributed to whatever is worthwhile in the reports that follow. Whatever isn't worthwhile is the result of my failure to take their advice.

I will soon turn to the information gathered in the survey. Before doing so, however, I want to recommend a general caution about interpreting it, namely, be careful about generalizing the survey findings to all CLLers or to the CLLers on the list who did not fill out and return a questionnaire.

The survey findings apply to those who participated in the survey. They may not apply to these others. Because survey participants were self-selected rather than randomly selected from the universe of CLLers, we cannot apply the usual statistical "tests of significance" that would tell us the chance that what's true about us is true about CLLers generally. Beyond that, information from the National Cancer Institute (some of which I shall cite later) shows that, as a group, some of our characteristics are quite different than those of CLLers as a whole. This may mean that still other things, true about us, are not true, or different for, CLLers as a whole. We have no good way of knowing.

Something like this is also the case with respect to the CLLers on the list who didn't participate in the survey. Those who participated, selected themselves. So we can't apply statistical tests of significance. Further, we don't know anything much about the roughly 55% of list subscribers who didn't participate--not even how many *are* CLLers! So, we can't be confident that our findings apply to them.

We can, however, be confident that our findings apply to the study group. As you'll soon see, many are of considerable interest. Or, more modestly, let me put it this way: I find them interesting. They're things I couldn't have known without a survey, and some seem to me helpful in looking at what may be ahead. I hope that you'll also find them of interest and help.

The first information from the survey will follow. It will outline some of the personal and social characteristics of the study group.

We asked survey participants to indicate age on their last birthday, gender, current marital status and living arrangements, geographic location, ethnic group, and occupation. Such matters are intrinsically interesting to most of us. They show us how we differ as a group from a slice of the general community. They can also give us a hint about how we differ as a group from all CLLers. Additionally, some of these items might provide clues to why experiences with CLL seem to vary so much.

In this report, I outline the findings on current age, gender, marital status, geographic location and ethnicity. The question about living arrangements resulted in many ambiguous answers, so I'm leaving it aside. I hope to report on occupations later; at the moment, they remain only partially classified. I conclude the report with a comment on some ways the characteristics of the study group probably differ from those we might find in a representative sample of all CLLers.

Study Group Characteristics

Table 1 shows the current ages of study group members. Underneath the table, I have shown the average current age of group members, their median age and the ages of the youngest and oldest.

A word about reading the table. (If you're used to reading tables, you can profitably skip this and the next paragraph.) The column on the far left, labeled "Current Age," shows how the individual ages are classified in the table. The middle column, "Number," shows the number of group members in each of these classes. Thus, there is one person under 20 years of age, five between 30 and 34, 71 between 50 and 54, and so on. If a class is missing, there was no person in that class; notice that nobody in the study group is currently 20 to 24 or 25 to 29 years of age. The last column, "Percent," shows the percentage of the group made up by each age class. You can see, for example, that people between the ages of 50 and 54 make up 20.9% of the group, while those between 75 and 79 make up 3.5%.

[The numbers beneath the first table show the mean current age of group members (what most of us call the "average"), their median current age, and the ages of the group's youngest and oldest members. You get the mean by dividing the sum of the ages by the number of people in the group. On the average, group members are a bit over 55. The median is the middle case: half of the group is younger and half older than 55. The youngest group member is 17; the oldest, 86.]

CLL is often said to be an old people's disease. But looking at the numbers in Table 1, I don't think one could fairly say that the survey respondents, as a group, are "old people." If not exactly persons in the first bloom of youth (or, as my mother used to say, "spring chickens"), we are hardly a group of old people either, at least by modern standards.

Two men for every woman. Is this the ratio we would find in a representative sample of all CLLers? Later in the report, I shall give a tentative answer to this question. Here, I want to note that, by and large, similar proportions of men and women fall in each age group; at most, there's a slight tendency for more of the men to be 65 or older (21% of the men versus 14% of the women). (Here and elsewhere, to conserve space and the reader's [and my] patience I have not included the tables showing these cross-tabulations. They are available on request.)

Most of us are currently married, and fewer than one in 20 study group members has never been married. Among those ever married, a larger proportion of women than men are divorced, separated or have lost a spouse (23% of the women versus 7% of the men). Finally, group members under age 40 are especially likely never to have been married; members age 50-59 are more likely to be divorced; and those age 70 and over are disproportionately likely to have lost a spouse.

One might speculate about the reasons for these relationships. Are husbands more likely to leave sick spouses? Does being diagnosed with CLL deter young people from marrying? Nothing in our data helps to explain these relations but they are worth thinking about, it seems to me, for what they might say about living with CLL.

Where do study group members live? Table 4 provides the answers.

The table shows the nine countries outside the USA in which group members reside. The bias toward English-speaking countries is obvious. The 291 USA respondents come from 43 states and the District of Columbia. California has the largest number of study group members, with almost twice as many (39) as the next highest state, New York (22). Other states with more than 10 study group members are Florida, Massachusetts, New Jersey, Illinois, Pennsylvania, Washington, Texas and Virginia.

Finally, a word about ethnicity. The study group is almost entirely Caucasian. One person is recorded as Asian, three as Hispanic. Three of the four non-Caucasians reside in the USA; one Hispanic lives in Bolivia.

In future reports, I shall look at some of the ways these differences may influence a person's CLL experience. Here, I want to comment briefly on what some of the characteristics of the study group suggest about its representativeness. The bottom line is that the study group is quite special.

The study group is obviously not a representative slice of the general population. There's also the strong possibility that the study group is quite unlike CLLers in general. My comments below are directed to this possibility because it follows that what we learn from the study group may need to be supplemented before it can be applied to the CLL population at large.

The grossest difference is in ethnicity. Consider the study group CLLers from the United States. There's not a single African-American. But data from the National Cancer Institute's SEER Program (Surveillance, Epidemiology, and End Results Program), and other sources, shows that a great many African-Americans have CLL; their diagnosis rates are apparently about two-thirds those of Caucasians. Apparently rates among Hispanics and Asians are lower, but they are hardly so low that a representative group the size of ours would have so few Hispanics and Asians among its members.

The CLL experiences of members of other than the Caucasian ethnic group may differ considerably from those of Caucasians. Unfortunately, we have too few in the group even to begin such an inquiry.

(The brother of an African-American CLLer did submit a questionnaire. He subscribes to the list to get information relevant to relating to his sister. He did not give us information about his sister. She could not thus be included in the study group.)

The study group also almost certainly differs in age from a representative group of CLLers. More specifically, by comparison with CLLers at large, I believe that the study group contains larger proportions of persons in the younger age categories, particularly in years 45-64. SEER data show that the chance of being diagnosed with CLL rises rapidly with age. Persons 65-69 years old are about 2-1/2 times as likely to be diagnosed with CLL as people aged 55-59. Those 75-79 years of age are 3-1/2 times as likely. Of course, there are fewer persons in the older age categories in the general population, so higher rates might still produce fewer CLLers in the upper age categories, resulting in something like the distribution shown in Table 1. (Think of earning a high rate of interest on a small amount of money versus a much lower rate on a much larger amount.) However, even after taking this into account, it seems to me reasonably certain that if we had a representative sample of CLLers at large, each higher age group would contain a larger number of CLLers, until ages 75-79, when the numbers would begin slowly to decrease. The numbers of study group members, on the other hand, increase to age 50-54 and then begin to decrease.

Putting the matter somewhat differently, while more than eight in ten study group members are under age 65, some rough and ready calculations (using diagnosis rates and census figures) suggest that fewer than one in three CLLers in the general population is that young. Said the other way around, two-thirds of the CLLers living in the community are 65 years of age or more, compared to less than one-fifth of those in the study group.

As we'll later see, even our limited data give us reason to think that younger and older people have different CLL experiences. So, another reason for caution in assuming that what we may learn about our group as a whole applies to all CLLers.

Finally, consider that the study group has two men for every woman. Is this the ratio we would find in a representative sample of all CLLers? It's difficult to be certain. Still, annual counts of newly diagnosed CLLers can be derived from NCI's SEER data. In those counts, the ratio tends to be roughly three men for every two women. My guess (and that's what it is) is that a representative sample of all CLLers would show something closer to a three to two gender ratio rather than the two to one ratio of the study group. This too limits the applicability of our findings.

How can we account for these kinds of differences? There's really nothing in the survey data that explains them. On the other hand, the 80% of study group members who answered the survey themselves subscribe to the CLL mailing list. This implies access to the Internet. And there are studies suggesting that people who use the Internet tend to be disproportionately Caucasian, young and male. My guess is that the greater proportions of people in these categories in the study group, compared to what is probably true of all CLLers, is mainly a consequence of such CLLers being more likely to use the Internet and subscribe to the CLL mailing list. (A straw of evidence supporting this interpretation is that the mean age of self-reporting study group members is 51, while the mean age of members whose questionnaires were sent in by helpers is nearly 56.)

There's certainly more to be said, but I won't try to say it here and now. Some of you may want to look at NCI's SEER data In the next report, we shall finally get to data not just on CLLers, but their experiences with CLL.

o CLL Survey, Table of Contents
o CLL Survey, page 2
o CLL Survey, page 3
o CLL Survey, page 4
o CLL Survey, page 5
o CLL Survey, page 6
o CLL Survey, page 7
o CLL List Homepage
o CLL FAQ Contents