Objective
We sought to estimate the frequency of self-reported fecal incontinence (FI), identify what proportion of these patients have a diagnosis of FI in their medical record, and compare health care costs and utilization in patients with different severities of FI to those without FI.
Study Design
Patients in a health maintenance organization were eligible and 1707 completed a survey. Patients with self-reported FI were assessed for a diagnosis of FI in their medical record for the last 5 years. Health care costs and utilization were obtained from claims data.
Results
FI was reported by 36.2% of primary care patients, but only 2.7% of patients with FI had a medical diagnosis. FI adversely affected quality of life and severe FI was associated with 55% higher health care costs (including 77% higher gastrointestinal-related health care costs) compared to continent patients.
Conclusion
Increased screening of FI is needed.
Fecal incontinence (FI), which is defined as the accidental loss of solid or liquid stool, affects a large proportion of the noninstitutionalized population of the United States. Estimates range from 2.2% to as high as 24%, with most estimates in the range of 7–12%. It has a significant impact on quality of life (QOL) that may include embarrassment, social stigmatization, depression, and anxiety. FI may also contribute to the decision to place older patients into assisted living facilities. However, despite its high prevalence, its deleterious impact on QOL, and the availability of effective medical and surgical treatments, surveys suggest that only 10% to 30% of people with FI have discussed this with their physician.
The direct costs associated with providing care for patients with FI are largely unknown. This is due in part to the variety of treatments, ranging from medical and behavioral approaches to surgical repair with very different costs, and the failure of most cost analyses to differentiate between FI and urinary incontinence.
The aims of this study were to: (1) estimate the prevalence of FI in primary care and its impact on QOL and health care costs; (2) determine what proportion of patients with self-reported FI have a medical diagnosis of FI; (3) identify patient characteristics that increase the likelihood that FI is reported to a physician and diagnosed; and (4) estimate health care costs associated with FI. It was our hypothesis that patients who self-report FI are often undiagnosed and that those who self-report FI utilize more total and more gastrointestinal (GI)-related health care dollars, make more health care visits, and have lower health-related QOL.
Materials and Methods
This study represents a secondary analysis of a larger study. Briefly, a large prospective study was performed to assess the impact of chronic constipation on QOL and health care costs at a health maintenance organization (HMO), Group Health Cooperative of Puget Sound (GHC), Seattle, WA, which serves >500,000 residents in the state of Washington. GHC’s patient population is similar to the general population of Seattle, WA, except that it underrepresents the high and low extremes of socioeconomic status. When compared to the US census, the study population underrepresents Hispanics and African Americans and overrepresents Asians.
All patients were required to be at least 18 years old; to have made at least 1 clinic visit (index visit) to a primary care provider at GHC from Sept. 1, 2004, through Dec. 31, 2005; and to have been enrolled at GHC for all of the previous 5 years. Gynecology clinics were included with other primary care clinics. Exclusions were a history of GI cancer or resection except for appendectomy or cholecystectomy. Two groups were recruited: 676 patients with a clinical diagnosis of constipation (564.0X) at their index visit, and a control group of 1031 patients who were matched by stratified sampling to the constipated patients with respect to age and sex but who did not receive a clinical diagnosis of constipation either at their index visit or at any time in the previous 5 years. Controls were matched to constipated patients using a stratified sampling frame to recruit a specified number in each age range of both male and female patients. This control group was unselected except for age and sex and the exclusion of a clinical diagnosis of constipation.
The study received institutional review board approval by both the University of North Carolina at Chapel Hill and GHC. Patients received a full description of the study with all elements of informed consent. Informed consent was inferred from completion and return of the questionnaires.
The survey included questions about the patient’s symptoms and history of constipation; the Rome III Diagnostic Questionnaire modules for functional constipation, irritable bowel syndrome (IBS), and bloating; demographic information; and the Short Form Health Survey (SF12). The SF12 is a generic health-related QOL scale whose 12 questions address self-perceptions of both physical and mental health. It is scored by summing items on physical and mental scales and transforming these into T-scores such that a score of 50 corresponds to the mean for the healthy adult population and 1 SD below the mean for healthy controls is equivalent to a T-score of 40. A difference of 5 U is considered the minimally clinically important difference.
Patients were classified as fecally continent or incontinent based on their response to the following question in the initial survey: “In the last 3 months, how often have you accidentally leaked liquid or solid stool?” Response options were: never, <1 day a month, 1 day a month, 2-3 days a month, 1 day a week, >1 day a week, or every day. Patients who answered “never” were classified as continent, those who answered “1 day a month” or less often were categorized as infrequent FI, and those who answered “2-3 days a month” or more often were categorized as frequent FI, similar to Bharucha et al.
Health care costs and health care utilization were calculated (for consenting patients) from claims data collected by GHC for the previous 5 years using previously described methods. Outliers were accounted for by averaging cost and utilization data for each patient over a 5-year period; 5-year means were the unit of analysis. Actual costs (not charges) were expressed in 2005 US dollars. The average number of inpatient days for 5 years was obtained from these claim data and used as an index of medical comorbidity. This unvalidated index is conceptually similar to the Charlson comorbidity index, with which it correlates.
Statistical analyses were performed using software (SPSS, Version 16.0, SPSS Inc, Chicago, IL). Fisher’s exact test was used for dichotomous data and Student t test was used for continuous variables when the comparison groups were of similar size; however, for comparison of the 15 patients with medical diagnosis of FI to the 535 FI patients without a medical diagnosis, Mann-Whitney U tests were employed. General linear modeling was used to test for mediation. A P value of < .05 was considered significant.
Results
The control cohort included 1003 analyzable patients after excluding 28 who had missing data for the FI question. The constipation cohort included 655 patients after excluding 21 with missing data for the FI question. Average age was 66.29 years for controls and 66.02 years for constipated patients; 68% of controls and 69% of constipated patients were female.
Table 1 shows that FI at least once in the past month was reported by 34.1% (342/1003 patients) in the control cohort and by 39.4% (258/655 patients) in the constipated cohort ( P = .074). The overall prevalence of FI was 36.2%. Frequent (≥2 times/mo) FI was reported by 10.7% of controls and 13.3% of constipated patients. Because there was no significant difference between those with a clinical diagnosis of constipation and those without, the 2 cohorts were pooled in subsequent analyses.
| Variable | No FI | FI <2/mo | FI ≥2/mo | Statistical significance |
|---|---|---|---|---|
| Clinical diagnosis of constipation | ||||
| No | 65.9% (n = 661) | 23.4% (n = 235) | 10.7% (n = 107) |
|
| Yes | 60.6% (n = 397) | 26.1% (n = 171) | 13.3% (n = 87) | |
| Rome criteria for IBS | ||||
| No | 69.7% (n = 739) | 22.4% (n = 237) | 7.9% (n = 84) |
|
| Yes | 53.1% (n = 312) | 28.6% (n = 168) | 18.4% (n = 108) | |
| Loose stools | ||||
| Never or sometimes | 67.2% (n = 1006) | 23.6% (n = 353) | 9.3% (n = 139) |
|
| Often or always | 31.0% (n = 48) | 34.2% (n = 53) | 34.8% (n = 54) | |
| Hard stools | ||||
| Never or sometimes | 65.5% (n = 791) | 23.8% (n = 288) | 10.7% (n = 129) |
|
| Often or always | 59.3% (n = 261) | 26.6% (n = 117) | 14.1% (n = 62) | |
Table 1 also shows that patients meeting the symptom criteria for IBS were significantly more likely to report FI compared to patients without IBS. Having loose stools “often” or “always” was highly predictive of frequent FI, and having hard stools “often” or “always” was weakly predictive of FI. The relationship between IBS diagnosis and FI was not explained by the effects of loose stools because, when analysis was restricted to patients who reported loose stools and hard stools “never” or “sometimes,” those with IBS were still more likely to report frequent FI (13.9% compared to 5.8% for non-IBS patients; χ 2 = 23.004; P < .001).
Demographic variables that have been reported to influence the development of FI are shown in Table 2 . FI was significantly associated with age, marital status, and race but not sex, education, or income.
| Variable | No FI | FI <2/mo | FI ≥2/mo | Statistical significance |
|---|---|---|---|---|
| Mean age, y | 63.2 | 69.6 | 73.6 |
|
| Sex | ||||
| Female | 53.6% (n = 719) | 23.2% (n = 260) | 12.7% (n = 142) |
|
| Male | 63.5% (n = 333) | 26.7% (n = 140) | 9.7% (n = 51) | |
| Marital status | ||||
| Single, separated, divorced, or widowed | 57.8% (n = 312) | 26.3% (n = 142) | 15.9% (n = 86) |
|
| Living with partner or married | 66.9% (n = 739) | 23.4% (n = 259) | 9.7% (n = 107) | |
| Education | ||||
| High school or less | 62.7% (n = 235) | 24.3% (n = 91) | 13.1% (n = 44) |
|
| College or professional | 64.4% (n = 816) | 24.4% (n = 309) | 11.2% (n = 142) | |
| Race | ||||
| Caucasian | 62.7% (n = 912) | 25.3% (n = 368) | 12.0% (n = 174) |
|
| Other races | 71.9% (n = 146) | 18.2% (n = 37) | 9.9% (n = 20) | |
| Personal income | ||||
| $0-$49,999 | 62.4% (n = 653) | 25.0% (n = 262) | 12.5% (n = 131) |
|
| $50,000-$99,000 | 68.1% (n = 246) | 23.5% (n = 85) | 8.3% (n = 30) | |
| ≥$100,000 | 68.5% (n = 37) | 20.4% (n = 11) | 11.1% (n = 6) | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
