On 2014 Jan 06, Tom Kindlon commented:

Why are we only given information on 3 out of the 8 SF-36 subscales?

Reading this paper, one could be forgiven for thinking that the SF-36 questionnaire only has 3 subscales: Physical Functioning, Mental Health and Social Functioning as that is all we are given information on. In fact, of course, the SF-36 questionnaire has 8 subscales: Physical function, Role physical, Bodily pain, General health, Vitality, Social function, Role emotional and Mental health[1]. The authors use the empiric definition for CFS[2] which requires at a minimum that the "role physical" and "role emotional" subscales also be measured. We also know that all 8 subscales were measured in this cohort[3]. So why was the information not given?

If one was not giving the authors the benefit of the doubt, one could speculate that it was because Table 4 would not look as good, as the Chi-squared calculations would not reach statistical significance for the missing data. But that would be speculation - there could be other reasons for the missing information.

Perhaps the authors could post the relevant data now.

I am not simply being mischievous - I would be interested in particular to see what are the scores for Classes 1 and 2 which include nearly all of the CFS patients (88/92, 95.7%).

References:

[1] Ware JE, Sherbourne CD: The MOS 36-item short form health survey (SF-36): conceptual framework and item selection. Med Care 1992, 30:473-483.

[2] Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C: Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Medicine 2005, 3:19.

[3] An evaluation of exclusionary medical/psychiatric conditions in the definition of chronic fatigue syndrome. Jones JF, Lin JM, Maloney EM, Boneva RS, Nater UM, Unger ER, Reeves WC. BMC Med. 2009 Oct 12;7(1):57. - see Tables 5 and 6.

On 2014 Jan 06, Tom Kindlon commented:

There has been criticism of how CFS is defined in this study

I thought it would be useful to point out that there is controversy [1,2] with regard to the criteria [3] used in this study to define Chronic Fatigue Syndrome (CFS).

For example, the criteria for CFS used in this study do not even require a patient to have fatigue. The authors say: “We used the Multidimensional Fatigue Inventory (MFI-20) [4] to measure characteristics of fatigue” but they do not give the thresholds. Given the MFI-20 has five subscales: (General fatigue, Physical fatigue, Mental fatigue, Activity reduction and Motivation reduction), one would probably suspect that a patient would have to score poorly on one of the headings which have fatigue in their title. But the actual criteria are: a patient needs to score >=13 on MFI general fatigue or >=10 on reduced activity. Note, one could score >=10 on the MFI reduced activity questions without necessarily being fatigued (one could be depressed or even lazy) (only current major depressive disorder with melancholic features (MDDm) is an exclusion for this definition of CFS).

This is despite the fact that in the current paper, the authors say: “Chronic fatigue syndrome (CFS) is a common, debilitating illness whose hallmark symptoms involve fatigue and fatigability”. Many other questions have been raised about the criteria for CFS that were used in this study. For example, the authors only considered current MDDm to be exclusionary for CFS while the International CFS Study group recommended that conditions (including MDDm) were considered exclusions unless they had been “resolved for more than 5 years before the onset of the current chronically fatiguing illness”[5].

Prevalence figures show that the criteria, that were used for this cohort, are selecting a broader group than previous criteria for CFS. Based on the figures derived from this cohort, the prevalence of CFS was estimated at 2.54% [6]. Other studies using similar methodology (but which did not operationalize the criteria [7] for the CFS in the same way as this study) estimated the prevalence of CFS to be 0.235% (95% confidence interval, 0.142%-0.327%) and 0.422% (95% confidence interval, 0.29%-0.56%) [8,9].

References:

[1]. Jason LA, & Richman JA. How science can stigmatize: The case of chronic fatigue syndrome. Journal of CFS 2007;14:85-103.

[2]. Jason LA, Najar N, Porter N, Reh C. Evaluating the Centers for Disease Control's empirical chronic fatigue syndrome case definition. Journal of Disability Policy Studies 2009;20;93.

[3]. Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C: Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Medicine 2005, 3:19.

[4]. Smets EM, Garssen B, Bonke B, De Haes JC. The multidimensional fatigue inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995; 39: 315–25.

[5]. Reeves WC, Lloyd A, Vernon SD, Klimas N, Jason LA, Bleijenberg G, Evengard B, White PD, Nisenbaum R, Unger ER; International Chronic Fatigue Syndrome Study Group. Identification of ambiguities in the 1994 chronic fatigue syndrome research case definition and recommendations for resolution. BMC Health Serv Res. 2003 Dec 31;3(1):25.

[6]. Reeves WC, Jones JF, Maloney E, Heim C, Hoaglin DC, Boneva RS, Morrissey M, Devlin R. Prevalence of chronic fatigue syndrome in metropolitan, urban, and rural Georgia. Popul Health Metr. 2007 Jun 8;5:5.

[7]. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome; a comprehensive approach to its definition and study. Ann Int Med 1994, 121:953-959.

[8]. Reyes M, Nisenbaum R, Hoaglin DC, Unger ER, Emmons C, Randall B, Stewart JA, Abbey S, Jones JF, Gantz N, Minden S, Reeves WC: Prevalence and incidence of chronic fatigue syndrome in Wichita, Kansas. Arch Int Med 2003, 163:1530-1536.

[9]. Jason LA, Richman JA, Rademaker AW, Jordan KM, Plioplys AV, Taylor RR, McCready W, Huang CF, Plioplys S. A community-based study of chronic fatigue syndrome. Arch Intern Med. 1999 Oct 11;159(18):2129-37.

On 2014 Jan 06, Tom Kindlon commented:

A more detailed comparison would need to be made before one could say this replicates the previous study

Part of the aim of this study [1] appears to be to compare the classes that were drawn up with a previous cohort[2-4]. However it does not, to my mind, deal with this in a particularly rigorous fashion. The main quantitative comparisons are the percentages that fall in each class (Tables 6 and 7). However, the percentages will be influenced by the quantity and type of non-CFS controls used which are not the same in each cohort [to explain why this is important using an extreme example: if there were 1000 non-CFS cases for everyone one CFS case in one cohort, the percentages in each class would be different than if there was a 1:1 ratio of CFS to non-CFS cases in the other cohort].

The first study involved the following[5]: "This population-based case control study enrolled 227 adults identified from the population of Wichita with: (1) CFS (n = 58); (2) non-fatigued controls matched to CFS on sex, race, age and body mass index (n = 55); (3) persons with medically unexplained fatigue not CFS, which we term ISF (n = 59); (4) CFS accompanied by melancholic depression (n = 27); and (5) ISF plus melancholic depression (n = 28)." This was based on the classification in 1997-2000. These were then assessed during 2003. As one can see in Table 2, in 2003, 6 out of the original 58 CFS patients satisfied the CFS definition[6] as originally operationalized, along with 4 out of the controls. 6 more who had previously been excluded because of previous diagnosis of Major Depressive Disorder with melancholic features (MDDm) were also said to satisfy the original CFS diagnosis. The method for operationalizing the CFS definition[6] was then changed so there was then 43 individuals with CFS (see Table 5). Although the same method[5] of operationalizing the CFS definition[6] is used when comparing the Wichita and Georgia cohorts, it is a very different way to select patients and controls than the current study[1]. So it is questionable how interesting it is to compare the percentages in each class.

A comparison of the percentages of CFS in each class might have been interesting but that was not done.

Also, apart from the percentages, no tables with quantitative information are presented in the current paper to help the reader compare the class groups to see how valid the comparisons are. This is made more difficult because the original study gave much more detailed data on the six class solution rather than the five class solution [2]: "As the five- and six-class solutions produced practically identical classes, with the exception of the fifth group in the five-class solution being divided into the fifth and sixth classes in the six-class solution, only the six-class solution is presented in Table 2."

So in that paper, one has classes which have a median BMI of 32, 30 and 30 which are described in the current paper[1] as obese classes while class 5 would be a combination of classes 5 and 6 which have a median BMI of 26 and 27 are classed as non-obese. So numerical comparisons would have been of more use rather than looking at verbal descriptions - describing two groups which have a median BMI of 30 as obese (so approx 50% would have a BMI under 30, one threshold for obesity) and another group which has a median BMI of around 26.5 as non-obese, seems a bit unsatisfactory. There is a 5 class LCA solution in Figure 1 in one of the Wichita papers which gives some verbal descriptions[4]. As one can see, "obese" is only used to describe two of the five LCA groups: - Obese, hypnoea (27.93%) - Obese, hypnoea and stressed (15.32%) - Interoception (16.22%) - Interoception, depression (19.82%) - Well (20.72%)

However, this does not seem to be the same five class solution for the Wichita cohort as the one described in this paper as the percentages don't match up.

References:

[1]. Aslakson E, Vollmer-Conna U, Reeves WC, White PD. Replication of an empirical approach to delineate the heterogeneity of chronic unexplained fatigue. Popul Health Metr. 2009 Oct 5;7:17.

[2]. Vollmer-Conna U, Aslakson E, White PD: An empirical delineation of the heterogeneity of chronic unexplained fatigue in women. Pharmacogenomics 2006, 7(3):355-364.

[3]. Aslakson E, Vollmer-Conna U, White PD. The validity of an empirical delineation of heterogeneity in chronic unexplained fatigue. Pharmacogenomics. 2006 Apr;7(3):365-73.

[4]. Carmel L, Efroni S, White PD, Aslakson E, Vollmer-Conna U, Rajeevan MS. Gene expression profile of empirically delineated classes of unexplained chronic fatigue. Pharmacogenomics. 2006 Apr;7(3):375-86.

[5]. Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C. Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Med. 2005 Dec 15;3:19.

[6]. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A: The chronic fatigue syndrome; a comprehensive approach to its definition and study. Ann Int Med 1994, 121:953-959.

On 2014 Jan 06, Tom Kindlon commented:

A more detailed comparison would need to be made before one could say this replicates the previous study

Part of the aim of this study [1] appears to be to compare the classes that were drawn up with a previous cohort[2-4]. However it does not, to my mind, deal with this in a particularly rigorous fashion. The main quantitative comparisons are the percentages that fall in each class (Tables 6 and 7). However, the percentages will be influenced by the quantity and type of non-CFS controls used which are not the same in each cohort [to explain why this is important using an extreme example: if there were 1000 non-CFS cases for everyone one CFS case in one cohort, the percentages in each class would be different than if there was a 1:1 ratio of CFS to non-CFS cases in the other cohort].

The first study involved the following[5]: "This population-based case control study enrolled 227 adults identified from the population of Wichita with: (1) CFS (n = 58); (2) non-fatigued controls matched to CFS on sex, race, age and body mass index (n = 55); (3) persons with medically unexplained fatigue not CFS, which we term ISF (n = 59); (4) CFS accompanied by melancholic depression (n = 27); and (5) ISF plus melancholic depression (n = 28)." This was based on the classification in 1997-2000. These were then assessed during 2003. As one can see in Table 2, in 2003, 6 out of the original 58 CFS patients satisfied the CFS definition[6] as originally operationalized, along with 4 out of the controls. 6 more who had previously been excluded because of previous diagnosis of Major Depressive Disorder with melancholic features (MDDm) were also said to satisfy the original CFS diagnosis. The method for operationalizing the CFS definition[6] was then changed so there was then 43 individuals with CFS (see Table 5). Although the same method[5] of operationalizing the CFS definition[6] is used when comparing the Wichita and Georgia cohorts, it is a very different way to select patients and controls than the current study[1]. So it is questionable how interesting it is to compare the percentages in each class.

A comparison of the percentages of CFS in each class might have been interesting but that was not done.

Also, apart from the percentages, no tables with quantitative information are presented in the current paper to help the reader compare the class groups to see how valid the comparisons are. This is made more difficult because the original study gave much more detailed data on the six class solution rather than the five class solution [2]: "As the five- and six-class solutions produced practically identical classes, with the exception of the fifth group in the five-class solution being divided into the fifth and sixth classes in the six-class solution, only the six-class solution is presented in Table 2."

So in that paper, one has classes which have a median BMI of 32, 30 and 30 which are described in the current paper[1] as obese classes while class 5 would be a combination of classes 5 and 6 which have a median BMI of 26 and 27 are classed as non-obese. So numerical comparisons would have been of more use rather than looking at verbal descriptions - describing two groups which have a median BMI of 30 as obese (so approx 50% would have a BMI under 30, one threshold for obesity) and another group which has a median BMI of around 26.5 as non-obese, seems a bit unsatisfactory. There is a 5 class LCA solution in Figure 1 in one of the Wichita papers which gives some verbal descriptions[4]. As one can see, "obese" is only used to describe two of the five LCA groups: - Obese, hypnoea (27.93%) - Obese, hypnoea and stressed (15.32%) - Interoception (16.22%) - Interoception, depression (19.82%) - Well (20.72%)

However, this does not seem to be the same five class solution for the Wichita cohort as the one described in this paper as the percentages don't match up.

References:

[1]. Aslakson E, Vollmer-Conna U, Reeves WC, White PD. Replication of an empirical approach to delineate the heterogeneity of chronic unexplained fatigue. Popul Health Metr. 2009 Oct 5;7:17.

[2]. Vollmer-Conna U, Aslakson E, White PD: An empirical delineation of the heterogeneity of chronic unexplained fatigue in women. Pharmacogenomics 2006, 7(3):355-364.

[3]. Aslakson E, Vollmer-Conna U, White PD. The validity of an empirical delineation of heterogeneity in chronic unexplained fatigue. Pharmacogenomics. 2006 Apr;7(3):365-73.

[4]. Carmel L, Efroni S, White PD, Aslakson E, Vollmer-Conna U, Rajeevan MS. Gene expression profile of empirically delineated classes of unexplained chronic fatigue. Pharmacogenomics. 2006 Apr;7(3):375-86.

[5]. Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C. Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Med. 2005 Dec 15;3:19.

[6]. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A: The chronic fatigue syndrome; a comprehensive approach to its definition and study. Ann Int Med 1994, 121:953-959.

On 2014 Jan 06, Tom Kindlon commented:

There has been criticism of how CFS is defined in this study

I thought it would be useful to point out that there is controversy [1,2] with regard to the criteria [3] used in this study to define Chronic Fatigue Syndrome (CFS).

For example, the criteria for CFS used in this study do not even require a patient to have fatigue. The authors say: “We used the Multidimensional Fatigue Inventory (MFI-20) [4] to measure characteristics of fatigue” but they do not give the thresholds. Given the MFI-20 has five subscales: (General fatigue, Physical fatigue, Mental fatigue, Activity reduction and Motivation reduction), one would probably suspect that a patient would have to score poorly on one of the headings which have fatigue in their title. But the actual criteria are: a patient needs to score >=13 on MFI general fatigue or >=10 on reduced activity. Note, one could score >=10 on the MFI reduced activity questions without necessarily being fatigued (one could be depressed or even lazy) (only current major depressive disorder with melancholic features (MDDm) is an exclusion for this definition of CFS).

This is despite the fact that in the current paper, the authors say: “Chronic fatigue syndrome (CFS) is a common, debilitating illness whose hallmark symptoms involve fatigue and fatigability”. Many other questions have been raised about the criteria for CFS that were used in this study. For example, the authors only considered current MDDm to be exclusionary for CFS while the International CFS Study group recommended that conditions (including MDDm) were considered exclusions unless they had been “resolved for more than 5 years before the onset of the current chronically fatiguing illness”[5].

Prevalence figures show that the criteria, that were used for this cohort, are selecting a broader group than previous criteria for CFS. Based on the figures derived from this cohort, the prevalence of CFS was estimated at 2.54% [6]. Other studies using similar methodology (but which did not operationalize the criteria [7] for the CFS in the same way as this study) estimated the prevalence of CFS to be 0.235% (95% confidence interval, 0.142%-0.327%) and 0.422% (95% confidence interval, 0.29%-0.56%) [8,9].

References:

[1]. Jason LA, & Richman JA. How science can stigmatize: The case of chronic fatigue syndrome. Journal of CFS 2007;14:85-103.

[2]. Jason LA, Najar N, Porter N, Reh C. Evaluating the Centers for Disease Control's empirical chronic fatigue syndrome case definition. Journal of Disability Policy Studies 2009;20;93.

[3]. Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C: Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Medicine 2005, 3:19.

[4]. Smets EM, Garssen B, Bonke B, De Haes JC. The multidimensional fatigue inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995; 39: 315–25.

[5]. Reeves WC, Lloyd A, Vernon SD, Klimas N, Jason LA, Bleijenberg G, Evengard B, White PD, Nisenbaum R, Unger ER; International Chronic Fatigue Syndrome Study Group. Identification of ambiguities in the 1994 chronic fatigue syndrome research case definition and recommendations for resolution. BMC Health Serv Res. 2003 Dec 31;3(1):25.

[6]. Reeves WC, Jones JF, Maloney E, Heim C, Hoaglin DC, Boneva RS, Morrissey M, Devlin R. Prevalence of chronic fatigue syndrome in metropolitan, urban, and rural Georgia. Popul Health Metr. 2007 Jun 8;5:5.

[7]. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome; a comprehensive approach to its definition and study. Ann Int Med 1994, 121:953-959.

[8]. Reyes M, Nisenbaum R, Hoaglin DC, Unger ER, Emmons C, Randall B, Stewart JA, Abbey S, Jones JF, Gantz N, Minden S, Reeves WC: Prevalence and incidence of chronic fatigue syndrome in Wichita, Kansas. Arch Int Med 2003, 163:1530-1536.

[9]. Jason LA, Richman JA, Rademaker AW, Jordan KM, Plioplys AV, Taylor RR, McCready W, Huang CF, Plioplys S. A community-based study of chronic fatigue syndrome. Arch Intern Med. 1999 Oct 11;159(18):2129-37.

On 2014 Jan 06, Tom Kindlon commented:

Why are we only given information on 3 out of the 8 SF-36 subscales?

Reading this paper, one could be forgiven for thinking that the SF-36 questionnaire only has 3 subscales: Physical Functioning, Mental Health and Social Functioning as that is all we are given information on. In fact, of course, the SF-36 questionnaire has 8 subscales: Physical function, Role physical, Bodily pain, General health, Vitality, Social function, Role emotional and Mental health[1]. The authors use the empiric definition for CFS[2] which requires at a minimum that the "role physical" and "role emotional" subscales also be measured. We also know that all 8 subscales were measured in this cohort[3]. So why was the information not given?

If one was not giving the authors the benefit of the doubt, one could speculate that it was because Table 4 would not look as good, as the Chi-squared calculations would not reach statistical significance for the missing data. But that would be speculation - there could be other reasons for the missing information.

Perhaps the authors could post the relevant data now.

I am not simply being mischievous - I would be interested in particular to see what are the scores for Classes 1 and 2 which include nearly all of the CFS patients (88/92, 95.7%).

References:

[1] Ware JE, Sherbourne CD: The MOS 36-item short form health survey (SF-36): conceptual framework and item selection. Med Care 1992, 30:473-483.

[2] Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C: Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Medicine 2005, 3:19.

[3] An evaluation of exclusionary medical/psychiatric conditions in the definition of chronic fatigue syndrome. Jones JF, Lin JM, Maloney EM, Boneva RS, Nater UM, Unger ER, Reeves WC. BMC Med. 2009 Oct 12;7(1):57. - see Tables 5 and 6.