As most of you are aware, the TMA has been administering a survey to its members since January, 1997. We have been entering data since the inception of the study. We continue to administer and collect surveys, and we encourage each of you to fill out the survey, if you have not already done so. All of the respondents to the surveys are being kept completely anonymous. There is no way to identify any of the responses in the analysis and results with a particular respondent. There are no names or identifying characteristics associated with any of the data being analyzed.

We are in the process of coding the survey responses, and beginning to perform some analysis of the data. What appears in this newsletter is a very brief report of some of the preliminary results. David and I would like to thank Dr. Hisham Choueiki for his assistance in performing our analysis and for reviewing our report.

It is critically important to keep in mind that our sample is not a randomly selected sample. Our sample is composed of individuals who have found The Transverse Myelitis Association and have chosen to complete and return the survey. In addition, there is a considerable response bias, which will have an impact on the results. For instance, most of our members have found us via the Internet. From what we understand of the digital divide, that likely means that our sample is under-represented by people who are minorities, low income and elderly. Additionally, it is likely, since the survey appears only in English, that our sample does not adequately represent people who do not have some facility with the English language. When we prepare the final report, we will offer a complete discussion of response bias. For the purposes of this article, it is merely important to consider that there is response bias.

Another important shortcoming of the data is that it is self-reported. We have no way of knowing whether all of the respondents received a good diagnosis of TM. The doctors who offered the TM diagnosis to our respondents used a set of criteria to make their diagnosis; we have no idea what those criteria were and we have no reason to believe that they all used the same criteria or that the criteria were consistently applied.

Having identified some of the serious problems with the data, I will close this introduction by noting that while there are certainly flaws in the data, it remains of extremely high value for the purpose of better understanding Transverse Myelitis. There are 546 respondents currently in our sample; there is no published research that comes close to this number of TM patients.

Table 1 presents the geographic distribution of the sample. The vast majority of the sample is from the United States; the UK and Canada represent the majority of the international respondents.

Table 2 identifies the current state residence of each of the respondents in our sample. The column "No. in Sample" reports this number. One of the issues we tested was whether any of the states were either over-represented or under-represented in our sample as compared to the state's proportionate population size. This test also addressed the representativeness of our sample as compared to the general population of the United States.

Table 1 Geographic Distribution of Sample

        Australia               7
        Brazil                  3
        Canada                  24
        Denmark                 2
        E Malaysia              1
        France                  1
        Germany                 1
        Greece                  1
        Holland                 1
        India                   1
        Ireland                 2
        Italy                   2
        Japan                   1
        New Zealand             1
        Republic of Korea       1
        Scotland                4
        South Africa            2
        The Netherlands         2
        Turkey                  1
        UK                      33
        USA                     455
        Total Sample            546

The United States population data was derived from the 2000 census. We generated two sets of percentages; the first was derived by dividing each state population by the total United States population and the second was derived by dividing the sample of TM respondents for each state by the total number in the sample. There were 51 pairs of percentages; we included the District of Columbia. While there are 455 US residents in our sample, only 454 were included in this analysis. One of the US respondents was serving on a military base and did not identify a state of residence on the survey.

We ran a paired T-test that compared each pair of percentages for each state. The Null hypothesis was the following: There is no difference between each pair of percentages; the percentage of TM respondents of each state versus the percentage of each state population. The alternative hypothesis states that there exists at least one state that is either over-represented or under-represented. The p -value from the T-test was 0.8042, which means that we failed to reject the Null hypothesis at the 0.05 level of significance.

State   State Pop       % State Pop     No. in Sample           % State in Sample       SampleN_Pop

AL      4447100         2% (0.0156)             8               2% (0.0176)             1.79893E-06
AK      626932          0% (0.0022)             2               0% (0.0044)             3.19014E-06
AZ      5130632         2% (0.0180)             9               2% (0.0198)             1.75417E-06
AR      2673400         1% (0.0094)             1               0% (0.0022)             3.74056E-07
CA      33871648        12% (0.1188)            46              10% (0.1013)            1.35807E-06
CO      4301261         2% (0.0151)             6               1% (0.0132)             1.39494E-06
CT      3405565         1% (0.0119)             3               1% (0.0066)             8.80911E-07
DE      783600          0% (0.0027)             1               0% (0.0022)             1.27616E-06
DC      572059          0% (0.0020)             2               0% (0.0044)             3.49614E-06
FL      15982378        6% (0.0560)             25              6% (0.0551)             1.56422E-06
GA      8186453         3% (0.0287)             15              3% (0.0330)             1.8323E-06
HA      1211537         0% (0.0042)             2               0% (0.0044)             1.6508E-06
ID      1293953         0% (0.0045)             2               0% (0.0044)             1.54565E-06
IL      12419293        4% (0.0435)             24              5% (0.0529)             1.93248E-06
IN      6080485         2% (0.0213)             9               2% (0.0198)             1.48015E-06
IA      2926324         1% (0.0103)             5               1% (0.0110)             1.70863E-06
KS      2688418         1% (0.0094)             2               0% (0.0044)             7.43932E-07
KY      4041769         1% (0.0142)             6               1% (0.0132)             1.4845E-06
LA      4468976         2% (0.0157)             7               2% (0.0154)             1.56635E-06
ME      1274923         0% (0.0045)             3               1% (0.0066)             2.35308E-06
MD      5296486         2% (0.0186)             14              3% (0.0308)             2.64326E-06
MA      6349097         2% (0.0223)             15              3% (0.0330)             2.36254E-06
MI      9938444         3% (0.0348)             16              4% (0.0352)             1.60991E-06
MN      4919479         2% (0.0172)             2               0% (0.0044)             4.06547E-07
MS      2844658         1% (0.0100)             4               1% (0.0088)             1.40614E-06
MO      5595211         2% (0.0196)             8               2% (0.0176)             1.42979E-06
MT      902195          0% (0.0032)             1               0% (0.0022)             1.10841E-06
NE      1711263         1% (0.0060)             3               1% (0.0066)             1.75309E-06
NV      1998257         1% (0.0070)             3               1% (0.0066)             1.50131E-06
NH      1235786         0% (0.0043)             3               1% (0.0066)             2.4276E-06
NJ      8414350         3% (0.0295)             11              2% (0.0242)             1.30729E-06
NM      1819046         1% (0.0064)             4               1% (0.0088)             2.19895E-06
NY      18976457        7% (0.0665)             27              6% (0.0595)             1.42282E-06
NC      8049313         3% (0.0282)             16              4% (0.0352)             1.98775E-06
ND      642200          0% (0.0023)             1               0% (0.0022)             1.55715E-06
OH      11353140        4% (0.0398)             27              6% (0.0595)             2.3782E-06
OK      3450654         1% (0.0121)             3               1% (0.0066)             8.694E-07
OR      3421399         1% (0.0120)             8               2% (0.0176)             2.33822E-06
PA      12281054        4% (0.0431)             28              6% (0.0617)             2.27993E-06
RI      1048319         0% (0.0037)             1               0% (0.0022)             9.53908E-07
SC      4012012         1% (0.0141)             9               2% (0.0198)             2.24326E-06
SD      754844          0% (0.0026)             2               0% (0.0044)             2.64955E-06
TN      5689283         2% (0.0199)             8               2% (0.0176)             1.40615E-06
TX      20851820        7% (0.0731)             21              5% (0.0463)             1.00711E-06
UT      2233169         1% (0.0078)             0               0% (0.0000)             0
VT      608827          0% (0.0021)             1               0% (0.0022)             1.6425E-06
VA      7078515         2% (0.0248)             12              3% (0.0264)             1.69527E-06
WA      5894121         2% (0.0207)             12              3% (0.0264)             2.03593E-06
WV      1808344         1% (0.0063)             3               1% (0.0066)             1.65898E-06
WI      5363675         2% (0.0188)             11              2% (0.0242)             2.05083E-06
WY      493782          0% (0.0017)             2               0% (0.0044)             4.05037E-06
Total    281,421,906                      Total: 454

In other words, our sample is representative of the United States population. Additionally, there are no states that are either over-represented or under-represented in the population. There are more people from the state of California with TM in our sample, because this state has the highest population of any state in the country. When comparing the percentage of people who live in each state to the percentage of people in our sample from each state, the similarities are remarkable. We know that there are communities that have disproportionately high numbers of people with TM; far higher than can be explained by their population size. From this analysis, however, there does not appear to be any higher frequency of TM in one state as compared to another that cannot be explained by the size of the population.

The next issue we analyzed concerns the frequency of TM by latitude in the United States. There is a higher prevalence of Multiple Sclerosis in the more northern latitudes as compared to the central and southern latitudes. We performed a very preliminary study to test whether there was a higher prevalence of TM in the more northern latitudes in the United States. This analysis was also based on the data, which appears in Table 2.

We divided the United States into three regions. We did not attempt to divide states in this analysis; the entire TM state population was counted in one of the three regions. The northern region is composed of people from the following states: WA, OR, ID, MT, WY, ND, SD, MN, WI, MI, NY, MA, RI, VT NH, ME and CT. The central region is composed of people from CA, NV, UT, CO, NE, KS, IA, MO, IL, IN KY, OH, WV, PA, VA, NJ, DE, and MD. The southern region is composed of people from AZ, NM, OK, TX, AR, MS, LA, TN, AL, GA, FL, NC, and SC. In order to make a valid comparison of the regions, we created a variable called SAMPLEN_POP which took the number of TM patients for each state and divided it by that state's population. This new variable removes the bias effect of the population factor, i.e., normalizes the data. The E identified in this variable stands for exponent; thus, E-06 represents 10 to the negative 6th power. We ran an Analysis of Variance (ANOVA) test using SAMPLEN_POP as the dependent variable and Northern Region, Central Region, Southern Region as the independent variables. There were 48 observations. Hawaii, Alaska and the District of Columbia were excluded from the analysis. In addition, we ran a Tukey test on the means.

The Null hypothesis was the following: There is no difference among the means of the three regions. The p-value from the F-test was 0.3587, which means that we failed to reject the Null hypothesis. In other words, there was no difference between regions. The results from the Tukey test indicated that there were no significant differences at the 0.05 level for the three combinations of paired means. Basically, this test supports the results of the ANOVA test. Consequently, from our preliminary analysis, there does not appear to be a relationship between the frequency of TM and latitude.

The remaining analyses are based on the entire TM sample of 546 respondents. Table 3 presents the age of onset of TM for each of the respondents. The largest numbers of people contracted TM between 35 and 49 years of age. There are 202 persons or 37% of the total sample represented by these three age groups. For those under the age of 20, there are two distinctively larger groups; for those who contracted TM under the age of one year, and for those between the ages of 11 and 15.

Table 3 Age of TM Onset

      Age at Onset                 Frequency

         0 to 1                       31
         1 to 5                        3
         6 to 10                      10
        11 to 15                      32
        16 to 20                      14
        21 to 25                      18
        26 to 29                      30
        30 to 34                      42
        35 to 39                      81
        40 to 44                      55
        45 to 49                      66
        50 to 54                      36
        55 to 59                      48
        60 to 64                      27
        65 to 69                      30
        70 to 74                      11
        75 to 79                       9
        80 to 84                       3

Table 4 presents the gender of the respondents.

Table 4 Gender

        Gender  Frequency
        Female        358
        Male          188

Females compose 65% of the sample, while males only compose 34% of the sample.

Table 5 presents the number of respondents who reported multiple episodes of TM. The survey did not ask respondents if they had experienced multiple episodes. The survey asked the respondent to identify when they had contracted TM. For six of the respondents, they identified multiple dates of onset which reflect multiple episodes. It is quite possible that if we had asked for whether people had experienced multiple episodes, these numbers would have been higher. It is possible that some people only reported the date of their first episode of TM. This is an area that will be given great attention in our next survey.

Table 5 Number of Episodes

Number of Episodes      Frequency
        1                     534
        2                       5
        3                       1

There were six people who reported multiple episodes of TM; five reported two episodes and one reported three episodes.

Table 6 presents the month in which the respondents contracted TM. We were interested in identifying whether TM onset might be seasonal or if there was a pattern of TM onset that might coincide with "flu season."

Table 6 Month of Onset

         Month Contracted TM    Frequency
             January                   60
             February                  49
             March                     43
             April                     41
             May                       54
             June                      49
             July                      35
             August                    43
             September                 45
             October                   38
             November                  38
             December                  43

There does not appear to be any particular pattern in the months in which people contract TM. The largest number of people contracted TM in January. The second highest number contracted TM in May. Table 7 groups the months by season in order to analyze the seasonal onset of TM.

Again, there does not appear to be any distinctive pattern in the results. The lowest numbers of people contracted TM during the fall season and the highest during the winter season. Of the 538 valid responses to this question, 121 or 22% contracted TM during the fall months, and 152 or 28% contracted TM during the winter months.

Table 7 Seasonal Onset of TM

        March           43
        April           41
        May             54
        Spring          138
        June            49
        July            35
        August          43
        Summer          127
        September       45
        October         38
        November        38
        Fall            121
        December        43
        January         60
        February        49
        Winter          152

We also analyzed the monthly and seasonal onset of TM among the children who contracted TM. We were looking for a similar pattern of onset that might coincide with a "flu season" for those children who could be in daycare facilities, a preschool or in schools.

Tables 8-12 present the children from five different age groupings and the month in which they contracted TM. We have also aggregated the data for each age grouping using the seasons as identified in Table 7.

Table 8 Month of Onset Age: Under 1 Year

                Frequency

January         5
February        2
March           2    Fall:      8
April           1    Winter:    9
May             3    Spring:    6
June            1    Summer:    7
July            4
August          2
September       2
October         3
November        3
December        2

Table 9 Month of Onset Age: 1 to 5

                Frequency

February        1
August          1
November        1

        Fall:    1
        Winter:  1
        Summer:  1

Table 10 Month of Onset Age: 6 to 10

                Frequency

February        2
March           2
May             1    Fall:      1
June            1    Winter:    3
July            2    Spring:    3
October         1    Summer:    3
December        1

Table 11 Month of Onset Age: 11 to 15

                Frequency
January         8
February        3
March           1
May             2    Fall:      6
June            1    Winter:    15
July            4    Spring:    3
August          3    Summer:    8
September       2
October         1
November        3
December        4

Table 12 Month of Onset Age: 16 to 20

                Frequency
February        2
March           3
April           2    Fall:      3
May             1    Winter:    3
July            1    Spring:    6
August          1    Summer:    2
November        3
December        1

We were unable to identify any pattern in the data regarding the month or season of onset of TM and the children's ages. The highest month of onset was in January for children who are under one year and between 11 and 15 years of age. The only other notable characteristic in the data is that there were 15 children between the ages of 11 and 15 who contracted TM during the winter months. From this preliminary analysis on the data, there does not appear to be relationship between a particular month or season and the onset of TM.

The last table presents the length of time the respondent had TM when they completed the survey. This number was calculated by the difference between the respondent's current age (when they completed the survey) and their age at the onset of TM.

Table 13  Length of Time Had TM

     Years      Frequency
        0       143
        1       138
        2       47
        3       36
        4       32
        5       19
        6       12
        7       11
        8       10
        9       5
        10      14
        11      8
        12      8
        13      4
        14      1
        15      2
        16      4
        17      4
        18      3
        20      2
        21      2
        23      4
        24      2
        25      6
        26      1
        27      1
        28      2
        29      2
        31      1
        33      1
        34      2
        35      1
        37      1
        38      1
        47      1
        51      1
        74      1

There were 533 respondents who provided valid responses to the questions regarding their current age and their age at the onset of TM. Of the 533 respondents, 143 or 27% had TM for less than a year when they completed the survey. There were 138 or 26% of the respondents who had TM for between 1 and 2 years when they completed the survey. This means that more than half of the respondent sample is composed of people who had TM for less than two years when they completed the survey. There were 396 or 74% of these respondents who had TM for less than five years when they completed the survey. When discussing response bias, this is, perhaps, one of the more significant considerations to keep in mind. Our sample is significantly over-represented by people who are recently diagnosed with TM, and under-represented by people who have had TM for longer than five years.

A thorough analysis of the TMA survey data will continue. The results will be reported to our membership and to the medical community. It is critically important that if you have not completed the survey, you should do so. You can find the survey on our web site at: http://www.myelitis.org/survey1.htm. Please do your part to help the medical community better understand TM!