The Fecal Incontinence Severity Index

Volume

42

Diseases of the

CozoN RecrvM

Number

12

DECEMBER 1999 ORIGINAL CONTRIBUTIONS

Patient and Surgeon Ranking of the Severity of Symptoms Associated with Fecal Incontinence The Fecal Incontinence Severity Index Todd H. Rockwood, Ph.D.,* James M. Church, M.D.,-~James W. Fleshman, M.D.,:} Robert L. Kane, M.D.,* Constantinos Mavrantonis, M.D.,$ Alan G. Thorson, M.D.,[] Steven D. Wexner, M.D.,q] Donna Bliss, R.N., Ph.D.,# Ann C. Lowry, M.D.,** From the *Clinical Outcomes Research Center, University of Minnesota, Minneapolis, Minnesota, tDepartment of Colon and Rectal Surgery, Cleveland Clinic, Cleveland, Ohio, }Division of Colon and Rectal Surgery, Barnes Jewish Hospital, St. Louis, Missouri, §Department of Colon and Rectal Surgery, Cleveland Clinic .Florida, Ft. Lauderdale, Florida, ]]Division of Colon and Rectal Surgery, Creighton University, Omaha, Nebraska, q[Division of Colon and Rectal Surgery, University of Minnesota, Minneapolis, Minnesota, #School of Nursing, University of Minnesota, Minneapolis, Minnesota, and **Department of Surgery, University of Minnesota, Minneapolis, Minnesota PURPOSE: The purpose of this research was to develop and evaluate a severity rating score for fecal incontinence, the Fecal Incontinence Severity Index. METHODS: The Fecal Incontinence Severity Index is based on a type × frequency matrix. The matrix includes four types of leakage commonly found in the fecal incontinent population: gas, mucus, and liquid and solid stool and five frequencies: one to three times per month, once per week, twice per week, once per day, and twice per day. The Fecal Incontinence Severity Index was developed using both colon and rectal surgeons and patient input for the specification of the weighting scores. RESULTS: Surgeons and patients had very similar weightings for each of the type × frequency combinations; significant differences occurred for only 3 of the 20 different weights. The Fecal Incontinence Severity Index score of a group of patients with fecal incontinence (N = 118) demonstrated significant correlations with three of the four scales found in a fecal incontinence quality-of-life scale. CONCLUSIONS: Evaluation of the Fecal Incontinence Sever-

ity Index indicates that the index is a tool that can be used to assess severity of fecal incontinence. Overall, patient and surgeon ratings of severity are similar, with minor differences associated with the accidental loss of solid stool. [Key words: Fecal incontinence; Health surveys; Outcome assessment (health care); Severity of illness index] Rockwood TH, Church JM, Fleshman J~', Kane RL, Mavrantonis C, Thorson AG, Wexner SD, Bliss D, LowIT AC. Patient and surgeon ranking of the severity of symptoms associated with fecal incontinence: the fecal incontinence severity index. Dis Colon Rectum 1999;42:1525-1532. A

f u n d a m e n t a l issue related to the successful study of o u t c o m e s is the identification of the severity

of a condition. Severity m e a s u r e s are i m p o r t a n t in establishing the c o m p a r a b i l i t y of patients in o r d e r to assess the effectiveness of alternative m e t h o d s of treatment. 1 The goal of this research was to d e v e l o p

Supported by a contract between the Universityof Minnesota Clinical Outcomes Research Center and The American Society of Colon and Rectal SurgeW and the Minnesota Coton and Rectal Foundation. No reprints are available.

a n d evaluate a severity i n d e x for fecal i n c o n t i n e n c e (FD. T h e f u n d a m e n t a l strength o f s u c h a tool is to allow for a s s e s s m e n t o f severity i n d e p e n d e n t o f direct clinical observations; thus, a s s e s s m e n t c a n b e d o n e at

1525

1526

ROCKWOOD ETAL

any time and provide a cost-effective means of monitoring the severity of a patient's status w-ithout direct clinician contact. For m a n y conditions, objective data may be used to assess severity. The DeMeester Score relies on data from 24-hour patient monitoring. 2 Although there are numerous physiologic measurements used for incontinence, none have b e e n shown to reflect severity or response to therapy accurately. 3 As a group patients with FI have lower manometric pressures than controls~ but there is significant overlap. In addition, a worsening clinical status is not always associated with decreasing pressures. 4, 5 Pudendal nerve terminal motor latency has not b e e n shown to correlate with manometric pressures or clinical status. 6 There is controversy over whether pudendal nerve terminal motor latency status is predictive of outcome after sphincter repair. 7 Anal ultrasound reliably detects sphincter defects. The presence of a defect correlates with manometric pressures but not clinical function. 8 Patient history is thus the best w a y to estimate severity of fecal incontinence. One approach would be to record each patient's experience descriptively, but that information is difficult to use for comparison. Numerous scoring or grading systems are present in the literature. 9-2° A n u m b e r of systems include only the consistency of the leakage, ignoring the frequency of occurrence. Those scales sacrifice discriminatory p o w e r for simplicity. Other scales mix historical data with data from physical examination or testing. 14, 15 The numerical values assigned each data point seem to have b e e n chosen arbitrarily. Other scales mix lifestyle issues with type and frequency.~7, ~9 Although quality of life (QOL) instruments are related to severity, they are designed to measure the impact of a given condition on a patient's life. Intuitively, the more severe the condition is, the more impact it will have on QOL. Therefore, the two measures should be correlated. However, QOL instruments measure different aspects of a patient's condition and should not be considered a direct indicator of severity, because they do not assess the same thing (i.e., the same level of severity can affect different patients in dissimilar ways). To our knowledge n o n e of the scales have b e e n tested or c o m p a r e d with other measures (convergent validity analysis). The lack of a standard, validated, severity measure makes comparison of patients and the outcomes of treatment modalities difficult. The goal of this research was to develop a severity measure for fecal incontinence, the Fecal Incontinence

Dis Colon Rectum, December 1999

Severity Index (FISI). The initial step was to explore surgeon and patient severity~rankings of various components of FI and to provide a logical basis for the assigned severity score. The basic c o m p o n e n t s of the FISI include the type of incontinence (gas, mucus, liquid, or solid) and the frequency of occurrence. Information on these topics is invariably obtained through a self-report, whether during a clinical assessment or a patient self-report done outside of the clinical setting. PATIENTS AND

METHODS

Study Design A type × frequency matrix was developed based on the components of fecal incontinence. The focus was on four types of incontinence: gas, mucus, and liquid and solid stool. The frequency dimension used five time frames: two or m o r e times per day, once per day, two or more times per week, once per week, and one to three times per month. Figure 1 A illustrates the matrix that was presented to the participants. Participants in the research were given the 20-cell table shown in Figure 1 A and instructed to rank the severity of the items relative to each other, assigning a "1" to the most severe cell in the table and a "20" to the least severe. The participants were instructed not to use tie scores The type × frequency- matrix was administered to both surgeons and patients. A total of 26 colon and rectal surgeons completed the form. All of these were conducted in person, in a focus-group type of setting. A total of 34 patients completed the form. Twelve of these were conducted through the mail and the additional 22 were completed by patients in a colon and rectal surgery clinic (Minneapolis, MN). Analysis of the two separate patient groups demonstrated no significant differences b e t w e e n their rankings; these data have b e e n pooled for the analysis. Using patient and surgeon ratings, separate severity weighting systems were developed. One was based on patient ratings and one on surgeon ratings. The mean value for each of the 20 ceils s h o w n in Figure 1 was calculated. This m e a n value provides the weighting of each type × frequency toward the overall severity score. (Table 1 gives the resultant cell-by-cell weighting scores for both patients and surgeons. Note that for calculation of the FISI scores, the original responses have b e e n reverse coded so that a higher score indicates greater severity, e.g., 1 = least severe condition and 20 = most severe condition.)

FECAL INCONTINENCESEVERITYINDEX

Vol. 42, No. 12

1527

A 2 or More Times a Day

Once a Day

2 or More Times a Week

Once a Week

1 to 3 Times A Month

Gas

Mucus Liquid Solid

B

a. Gas

2 or More Times a Day

Once a Day

2 or More Times a Week

Once a Week

1 to 3 Times A Month

Never

[]

[]

[]

[]

[]

[]

.............................................................................................................................................................................................................

b. Mucus

[]

[]

[]

[]

[]

[]

c. Liquid Stool

[]

[]

[]

[]

[]

[]

d Solid Stool

[]

[]

[]

[]

[]

[]

Figure 1. Fecal Incontinence Severity Index. A, Event × frequency matrix presented to surgeons and patients to develop weightings and overall severity score. Participants were instructed to rank the importance of each cell by placing a "1" in the most severe cell and a "20" in the least severe cell. B. Fecal Incontinence Severity Index Question. Presented to the fecal incontinence study population, the question asked, "For each of the following, please indicate on average how often in the past month you experienced any amount of accidental bowel leakage: (Check only one box

per row.)" Table 1.

Surgeon and Patient Ratings of Fecal Incontinence Two or More Times per Day

Gas Mucus Liquid Solid

Once per Day

Two or More Times per Week

Once per Week

One to Three Times per Month

Patient

Surgeon

Patient

Surgeon

Patient

Surgeon

Patient

Surgeon

Patient

Surgeon

12/5.7 12/5.6 19/1.9 18/2.7

9/4.7 11/3.8 18/3.6 19/3.8

11/4.6 10/4.6 17/2.4 16/2.4

8/4.5 9/3.2 16/2.8 17/3.4

8/3.9 7/3.8 13/3.3 13/2.1"

6/4.6 7/2.1 14/2.7 16/3.2

6/3.2 5/3.1 10/3.7 10/2.7"

4/4.9 7/3.7 13/3.0 14/2.9

4/3.3 3/3.0 8/4.1 8/3.1"

2/3.9 5/4.6 10/4.1 11/3.7

Figures are mean/standard deviation. Note for calculation of the Fecal Incontinence Severity Index scores that the original responses have been reverse coded so that a higher score indicates greater severity, e.g., 1 = least severe condition and 20 = most severe condition. * Indicates significant difference at the .01 level (Bonferroni adjusted t-test of means).

The following is an example illustrating h o w the FISI is calculated: Patient A reported the following: gas once per day, mucus leakage never, liquid stool loss once per week, and solid stool loss one to three times per month. Taking the surgeon weighting scores from Table 1, their FISI score would be determined in the following

manner: gas once per day, +8; mucus leakage never, +0; liquid stool loss once per week, +13; solid stool loss 1-3 times per month, +11; for a total FISI score of 32. Using the patient ratings from Table 1, the FISI score would be: gas once a day, +11; mucus leakage never, ÷0; liquid stool loss once a week, +10; solid stool loss 1-3 times a month, +8; for a total FISI score of 29.

1528

ROCKWOOD ETAL

To evaluate the FISI, data from another study on an FI patient population was also used. In that study 118 people with FI were surveyed and patients reported the frequency of incontinence they were experiencing. 21 This survey included the question presented in Figure 1 B. The data on the frequency of occurrence for each of the types of leakage given in this self-report were used to calculate each patient's severity score. Another data source from this study that will be used is the quality-of-life scales in the Fecal Incontinence Quality of Life scale (FIQL).is These scales were used to test how well the FISI scores correlated with measures of fecal incontinence--specific qualityof-life measures. Analysis Evaluation of the severity rankings used two separate analyses. The first directly compared the weightings of the surgeons and patients to determine whether there were differences between them in their ranking of the severity of Ft. Two approaches were used for this analysis. First, a Pearson correlation examined the extent of agreement in the rankings between patients and surgeons. Second, the mean ranking of each cell in the table presented in Figure 1 A was compared using a t-test of means. 70

Dis Colon Rectum, December 1999

The second analysis sought to evaluate whether or not the rankings of one group (patient or surgeons) better predicts patient QOL by comparing the Pearson correlation of both the patient and surgeon rankings with each of the four FIQL scale scores. Finally, the sensitivity of using one or the other ranking was tested by comparing the scores generated for hypothetical patients. RESULTS The rankings for the surgeons and patients correlate very highly (r = 0.97; Fig. 2). Nonetheless, the surgeons and patients differ in the relative importance they place on some elements used to calculate the FISI score. Table 1 presents the results of the analysis that compares the mean rankings assigned to each of the cells by the surgeons and patients. Overall, patients and surgeons demonstrated consistency- in their severity rankings. As shown in Table 1, the weighting given by patients and surgeons were significantly different in only three of the twenty cells (Bonferroni adjusted 1-test of means). Surgeons assigned more weight (severity) to infrequent episodes of solid stool incontinence than patients. Because the rankings by each group for liquid and solid stools and for gas and mucus were essentially

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R2 = 0,9663

60

50

tO

40

GO

30

20 0

10

i

0

10

20

i

30

40

i

50

Patient

Figure 2. Comparison of patient and surgeon severity ratings.

6O

70

Vol. 42, No. 12

FECAL INCONTINENCE SEVERITYINDEX

the same, in seemed reasonable to collapse these into two types of discharge: liquid or solid stool and gas or mucus. Figure 3 compares the rankings for the collapsed event types. There are no significant differences for the Gas/Mucus combinations, but as expected three of the Liquid/Stool combinations demonstrate significant differences: one to three times per month, once per week, and twice per week (Bonferroni adjusted t-test of means).

1529

direction, did not demonstrate significant correlations with the severity scores.

DISCUSSION The goal of this research was to evaluate a scheme to assess the severity of FI. The findings from this preliminary research reveal a general pattern of similarities with a few differences in h o w patients and surgeons rank the severity. The primary difference occurs in the ranking of infrequent leakage of solid stool, which the surgeons rank as being more severe. Although these differences do exist, there is no significant gain from using one set of rankings over the other in the prediction of a patient's self-reported QOL. The fact that the two respondent groups agreed in the majority of cases is encouraging. The discrepancies reflect the interpretation each group places on the different types of incontinence. A real difference in perception is shown by the surgeons putting greater importance on incontinence of solid stool than patients did. The surgeons are more likely to

Correlation of Severity and Quality of Life The correlation between each of the four FIQL scale scores and the severity weights generated from each of the two sources (patient and surgeons) for the weightings is shown in Table 2. The correlations between the severity weights and the scales from the FIQL are similar for patients and surgeons for each of the four scales. Three of the scales demonstrate significant correlations with both the surgeon and patient severity scores: Lifestyle, Coping/Behavior, and Embarrassment ( P < 0.05). The Depression/Self-Perception scale, although correlated in the expected

Gas/Mucus 1-3x Month Gas/Mucus lx Week

~ l,,t~,t/

Gas/Mucus 2x Week

~

I? ~,,,:~~ '

~:t

~Bi~

[:J ~,tmH~;~ ~{~Btttt~@g@t

Gas/Mucus lx Day

~1!,~

t~IFII~I/@

*Liquid/Solid 1-3x Month ~I~'~,~:~@~

~#~@~t~~

*Liquid/Solid 2x Week Liquid/Solid lx Day

~

Liquid/Solid 2x Day

] [] Patient • Surgeon I

~

t~M'~!ttt~ I

0 Less

2

~

~'~ .......

~'~

I

I

4

6

.

.

.

.

~!~'~ l"

I

8

10

t~t~t ~ . ~ I

12

~

,, @ ~ % q ~ I

14

I

16

18

20 More

Severity Figure 3. Comparison of surgeon and patient ratings of fecal incontinence. Combining Gas/Mucus and Liquid/Solid categories, higher score indicates greater severity. * = significant difference, Bonferroni adjusted t-test of means.

ROCKWOOD E T A L

1530

Dis Colon Rectum, December 1999

Table 2. Surgeon and Patient Severity Ratings Correlation with Fecal Incontinence Quality of Life (FiQL) Scales Rating

Lifestyle

Coping/ Behavior

Depression/ Self-Perception

Embarrassment

Patient -.45~ -.29" -.20 -.381" Surgeon -.44~ -.32* -.23 -.391" Figures are Pearson correlation coefficients. Note that severity score and FIQL scale scoring run in opposite directions; a higher rating indicates more severe fecal incontinence, whereas for the FiQL scales a higher score indicates a higher functional status or quality of life. * Significant at P < .05. 1 Significant at P < .01.

emphasize a physiological interpretation of events, whereas patients are more conscious of leakage that can affect personal hygiene and provoke social embarrassment. Surgeons view solid stool loss as reflections of sphincter integrity and the adequacy of surgical repairs. The question then becomes how- important the differences are between the two groups in producing different levels of overall severity. As noted earlier, either weighting source produces scores that are significantly correlated with three of the four QOL measuresY The score for an individual patient might vary depending on whether surgeon or patient rankings were used. The extent of the effect of using different weighting sources at a patient level can be appreciated from the following example: Patient 1 reported the following: gas two or more times per day (patient weight, 12; surgeon weight, 9), mucus two or more times per week (patient weight, 7; surgeon weight, 7), solid stools once per day (patient weight, 16; surgeon weight, 17); thus, Patient l's severity score based on the patient rankings is 35 and based on the surgeon rankings is 33. Patient 2 had gas one to three times per month (patient weight, 4; surgeon weight, 2), mucus once per week (patient weight, 5; surgeon weight, 7), liquid stools once per w e e k (patient weight, 10; surgeon weight, 12), and no solid stool loss; the severity score based on patient weightings would be 19 and based on surgeon weightings would be 22. Although the differences between the two groups (patients and surgeons) are not great, they can produce slightly different results. The magnitude of the difference will depend on the underlying frequency of the type of FI problem in a given sample of patients. Patients able to control stool but not flatus or mucus will receive higher scores w h e n the patient ratings are used. It is unclear how important these differences are clinically. This pattern of modest differences in the impor-

tance assigned by professionals and consumers has been seen in other contexts. In a study examining the relative importance ratings of activities of daily living, professionals tended to assign higher value to more severe elements, whereas consumers placed higher importance on the more common events. 22 In the end, the choice of which rating group to use as the criterion standard may d e p e n d on what outcome is being emphasized. In the context of an evaluation of surgical success, the surgeon weightings make more sense, because they reflect sphincter competence. Gas and mucus production are less likely to be sensitive to surgical intervention. However, from the perspective of patient satisfaction with the result, these elements have high salience. Further research is necessary to confirm these findings and test both the surgeon and patient weightings in clinical settings. If the significant differences between patients and surgeons persist, ultimately a choice of which perspective to use as a standard will need to be made or both scores will need to be reported. Further work is also necessary to unde> stand whether combining the categories of gas and mucus and of liquid and solid stool is useful. Finally, the details of assigning specific weighting scores need further resolution.

REFERENCES 1. Smith M. Severity. In: Kane RL, ed, Understanding health care outcomes research. Gaithersburg: Aspen Publishers, 1997:129-52. 2. Jamieson JR, Stein HJ, DeMeester TR, et al. Ambulatory 24-h esophageal pH monitoring: normal values, optima/ thresholds, specificity, sensitivity, and reproducibility [see comments]. Am J Gastroenterol 1992;87:1102-11. 3. Shelton A, Madoff R. Defining anal incontinence: establishing a uniform continence scale. Semin Colon Rectal Surg 1997;8:54-60. 4. Felt-Bersma RJ, Klinkenberg-Knol EC, Meuwissen SG.

Vol. 42, No. 12

5.

6.

7.

8. 9.

10. 11.

12.

13.

14. 15.

16.

17.

18.

19. 20.

21.

22.

FECAL INCONTINENCE SEVERITYINDEX

Anorectal function investigations in incontinent and continent patients. Differences and discriminatory value. Dis Colon Rectum 1990;33:479-85; discussion 485-6. Rogers, J, Henry MM, Misiewicz JJ. Combined sensory and motor deficit in primary neuropathic faecal incontinence. Gut 1988;29:5-9. Tjandra J, Sharma B, McKirdy H, et al. Anorectal physioloM testing in defecatory disorders: a prospective study. Aust N Z J Surg 1994;64:322-6. Gilliland R, Altomare DF, tvioreira H Jr, et al. Pudendal neuropathy is predictive of failure following anterior overlapping sphincteroplasty. Dis Colon Rectum 1998; 41:1516-22. Bartram C, Law P. Anal endosonography: technique and application. Adv Gastrointest Radiol 1991; 1:101-15. Broddn G, Dolk A, HolmstrOm B. Recovery of the internal anal sphincter following rectopexy: a possible explanation for continence improvement. IntJ Colorectal Dis 1988;3:23-8. Corman ML. Gracilis muscle transposition for anal incontinence: late results. BrJ Surg 1985;72(Suppl):S21-2. Rainey, JB, Donaldson DR, Thomson JP. Postanal repair: which patients derive most benefit? J R Coil Surg Edinb 1990;35:101-5. Rudd ~GV. The transanal anastomosis: a sphinctersaving operation with improved continence. Dis Colon Rectum 1979;22:102-5. Womack NR, Morrison JF, Williams NS. Prospective study of the effects of postanal repair in neurogenic faecal incontinence. Br J Surg 1988;75:48-52. Kelly JH. Cine radiography in anorectal malformations. J Pediatr Surg 1969;4:538-46. Holschneider AM. Treatment and functional results of anorectal continence in children with imperforate anus. Acta Chit Belg 1983;82:191-204. Lunniss PJ, Kamm ~vL&,Phillips RK. Factors affecting continence after surgery for anal fistula. Br J Surg 1382; 81:1382-5. Rothenberger D. Anal incontinence. In: CameronJL, ed. Current surgical therapy. Philadelphia: Decker Mosby, 1984:185-94. Miller R, Bartolo DC, Locke-Edmunds JC, Mortensen NJ. Prospective study of conservative and operative treatment for faecal incontinence. Br J Surg 1988;75:101-5. Jorge JM, Wexner SD. Etiology and management of" fecal incontinence. Dis Colon Rectum 1993;36:77-97. Pescatori M, Anastasio G, Bottini C, Mentasti A. Newt grading and scoring for anal incontinence: evaluation of 335 patients. Dis Colon Rectum 1992;35:482-7. Rockwood TH, ChurchJM, Fleshman JW, et al. FIQL: a quality of life instrument for patients with fecal incontinence. Dis Colon Rectum (in press). Kane R, Rockwood T, Finch M, Philp I. Consumer and professional ratings of the importance of functional

1531

status components. Health Care Financing Rev 1997;19: 11-22.

Invited Editorial To the Editor--The study b y Rockwood and coworkers is a w e l c o m e addition to the literature on the quantitative assessment of fecal incontinence using a scoring system. There is no doubting the n e e d for an internationally agreed scoring system for fecal incontinence to allow- comparisons b e t w e e n patient p o p u lations and outcomes of medical and surgical treatment. Such a goal has b e e n achieved for urinary incontinence through the International Continence Society, but this b o d y is mainly the preserve of urologists and urogynecologists. A similar grouping with representatives from colorectal surgery, gastroenterology and associated specialists would be helpful. Agreeing a universal scoring system should be possible, but from w h o s e perspective should the system be created, the patients' or the surgeons'? Rockwood and colleagues first of all constructed a chart with four types of incontinence and five frequencies, from twice per day up to one to three times per month. The 20 cells could then be weighted 1 to 20 b y a group of patients and colorectal surgeons. Slightly confusingly, they adopted the convention of the lowest score "1" for the most severe symptoms and the highest "20" for the least severe. Although this score has b e e n inverted to make the highest value the worst incontinence in the final calculation of severity, the patients participating in the original ranking exercise might have found this off-putting. The m e a n score for each cell was then calo.llated for patients and surgeons, and on the whole there was a g o o d correlation. Interestingly, patients were m o r e concerned about frequent episodes of gas incontinence, whereas surgeons ranked less frequent episodes of solid stool more severely. It is a pity that the authors did not look at urgency in a similar way. Patients report urgency and the uncertainty that accompanies it an important and troublesome s y m p t o m in ulcerative colitis and incontinence. They then went on to compare their n e w Fecal Incontinence Severity Index with an as yet unpublished but validated Fecal Incontinence Quality of Life scale, and there was a similarly g o o d correlation. In discussion they did not r e c o m m e n d which rankings should be used, because it is d e a r that m u c h d e p e n d s on what is being a s s e s s e d - - s u r g e o n s look with the results of surgery in mind, are less able to achieve