| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Nuffield Department of Obstetrics and Gynaecology and Health Services Research Unit, Division of Public Health and Primary Health Care, University of Oxford and National Endometriosis Society, United Kingdom, and Searle Pharmaceuticals, United States, Chicago, Illinois.
Address reprint requests to: Georgina Jones, MA, Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Level 3, Women’s Centre, Oxford OX3 9DU, United Kingdom; E-mail: georgina.jones{at}obs-gyn.ox.ac.uk.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: The Endometriosis Health Profile-30 was developed in three stages. Stage 1 included open-ended exploratory interviews with 25 women to generate the items on the questionnaire. Stage 2 was an 87-item questionnaire administered in a postal survey to identify the most salient dimensions of health-related quality of life. In stage 3, the reliability and validity of the questionnaire were evaluated.
RESULTS: The final instrument contained a core questionnaire with 30 items and five scales: pain, control and powerlessness, emotional well-being, social support, and self-image. Six modular parts consisting of 23 questions were also developed and measured the areas of sexual intercourse, work, relationship with children, feelings about the medical profession, treatment, and infertility. All the scales achieved high internal reliability, with Cronbach’s 
coefficient ranging from 0.83 to 0.93 (core questionnaire) and 0.79 to 0.96 (modules). The intraclass correlation coefficients to evaluate the test-retest reliability were high (range 0.88–0.98, P < .001). Content validity was demonstrated as the questionnaire was developed from interviews of patients rather than existing literature and clinical scales. Construct validity was assessed by correlating the Endometriosis Health Profile-30 scales with the relevant Short Form-36 scales. High correlations for all comparisons were found (-0.41 to -0.73).
CONCLUSION: The Endometriosis Health Profile-30 is a reliable, valid, patient-generated instrument to measure the health-related quality of life of women with endometriosis. Its application in various health care settings will provide new and valuable information on the effect of endometriosis on health-related quality of life from the patients’ perspective.
Endometriosis is a chronic disease and one of the most common gynecologic conditions. The symptoms typically associated with endometriosis are a major cause of morbidity and psychosocial problems, but the mechanism is uncertain. The pelvic pain associated with endometriosis may lead to greater social dysfunction,1 although limited treatment options, associated infertility, and a delay in diagnosis can also lead to feelings of frustration and isolation. Despite a probable effect on the social and psychologic parameters, the impact of endometriosis on the patient’s health-related quality of life has been poorly researched. Health-related quality of life is a multidimensional concept encompassing physical, psychologic, and social aspects associated with a particular disease or its treatment.
A small number of studies have reported on the impact of endometriosis on health-related quality of life. However, only five of these have used instruments with established psychometric properties.2–6 Three generic measures have been used, the Nottingham Health Profile,2 Short Form-36,3,5 and Short Form-12.4 One limitation of using generic measures is that they may not be sensitive enough to assess changes in specific illnesses, as they were designed to measure health status across a wide variety of diseases.7 Disease-specific questionnaires that contain items developed from appropriate patient groups should be more responsive to changes in health status.8 Two disease-specific questionnaires for endometriosis have been developed.3,6 However, most of the items on the instruments were not derived from patients with the condition. There is now good evidence that the assessments patients make about their health differ from the proxy reports other health professionals make about their patients’ well-being.9
The aim of this research was therefore to design a disease-specific health-related quality-of-life questionnaire using established psychometric tests10 and based on interviews with women with endometriosis.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Individual, in-depth interviews were conducted with 25 women to explore the effects of endometriosis on their quality of life (group A). This sample size was determined at the point at which no new themes emerged.11 A pilot study (group B) checked the face validity of the items generated. No modifications were made because all patients were satisfied with the questionnaire.
The 87-item questionnaire was administered by mail to 1000 women requesting information about the National Endometriosis Society for the first time (group C). This sample size was adequate for factor analysis.12
Psychometric tests were carried out on the 87 items to identify the most salient dimensions of endometriosis that affect the patient’s quality of life. The extraction method used was principal component analysis and the rotation method used was varimax analysis. The principal component analysis is used to reduce numerous items into a smaller number of dimensions by statistically determining which items are related to others. This can then be analyzed using varimax rotation, which maximizes the amount of variance explained.12 The Cronbach statistic was used to measure the internal reliability; high scores usually indicate that scale items are measuring related constructs.13
Each scale was transformed on a range from 0 (indicating the best health status) to 100 (worst health status), enabling the extent of ill health to be measured (scale score = total of raw scores for each item in the scale/maximum possible raw score x 100). Test-retest reliability and construct validity were assessed by asking group D to complete one questionnaire the day it was received (time 1) and the second 3–6 days later (time 2). At the time of the second questionnaire, each respondent was asked to report any important health or life changes that had occurred between completing the questionnaires. The Wilcoxon signed rank test (nonparametric) was used to calculate the statistical differences between the scores at times 1 and 2. The intraclass correlation coefficient was calculated to examine the relationship between the scale scores at times 1 and 2.
To assess the construct validity, the disease-specific questionnaire and Short Form-36 were administered in a postal survey (n = 40) (group E). We hypothesized a significant correlation (Spearman 
nonparametric coefficient) would be found between scales with similar content on the Endometriosis Health Profile-30 (EHP-30) and Short Form-36. All statistical analyses were performed using the Scientific Package for Social Sciences 9.0 for Windows (Oxford University, United Kingdom).
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Of the 1000 questionnaires administered between October 1999 and March 2000, 371 (37.1%) were returned. Of these, 88% (n = 325) were suitable for inclusion in the study; 9% (n = 35) were returned by women who did not yet have surgical confirmation of the disease and 11 women (3%) no longer had symptoms and so were unable to complete the questionnaire. The mean age of the sample was 32.5 ± 7.2 years (range 17–58). The mean time since the diagnosis was 32 ± 44 months (range 1–360) (n = 323).
All questions answered without a "not relevant category" (n = 54) were selected for the scale generation and item reduction of the core questionnaire. The data were factor analyzed (varimax rotation). Only factors that gained an Eigen value (raw sum of the squares) of 1 or more were retained. Eleven factors were identified, accounting for 67.2% of the variance. However, only the first eight factors, accounting for 57.2% of the variance, were meaningful. Questions that obtained a value of 0.50 or more on any of the factors were retained. These eight factors included 41 questions that addressed pain (nine items); control and powerlessness (eight items); temperament (three items); social support (seven items); activities (four items); self-image (four items); emotional well-being (four items), and cognition (two items).
The internal reliability consistency of the scales was assessed using the Cronbach .13 Items were examined to check for ambiguity and similarity, resulting in the removal of one item from factor 2, two questions from factor 4, and one question from a four-item factor that addressed limitations to activities. The Cronbach statistic was calculated again on the 37 items, and the internal consistency reliability of the scales was unaffected. The 37 items were factor analyzed again to check the dimensions from the first factor analysis. No changes were found to the scales relating to social support, self-image, powerlessness, and control. However, four scales (pain and activities and emotional well-being and temperament) combined to produce two dimensions. This produced six factors with 34 items greater than 0.5, which addressed pain (11 items); control and powerlessness (six items); emotional well-being (six items); social support (five items); self-image (four items); and cognition (two items).
The Cronbach coefficient was calculated. The cognition scale was removed because the Cronbach at 0.69 was considered too low.14 Two additional items were removed on the basis of the low results of the corrected item to total correlations. The internal consistency was then reassessed. These 30 items were factor analyzed for a third time and remained the same. This resulted in a 30-item core questionnaire (the EHP-30) with five scales (Table 1
). Each scale was standardized on a scale ranging from 0 to 100; the descriptive statistics for the five dimensions are shown in Table 2.
|
|
was unaffected. These 23 items were factor analyzed again; all six scales stayed the same (Table 3). Each scale was then standardized on a scale of 0–100. Table 4 shows the descriptive statistics for the modular questionnaire. Table 5 presents the mean and confidence interval for both the core and modular scales.
|
|
|
shows the correlations of the four scales of the EHP-30 core questionnaire with the scales of the Short Form-36. Of the 83 questionnaires administered to assess the test-retest reliability, 54 (65%) were returned. Thirty-one women (57.4%) reported no change in their health status. No significant differences between the scores at times 1 and 2 were found on four of the five scales (EHP-30 pain, z = -0.949, P > .05; EHP-30 control and powerlessness, z = -1.483, P > .05; EHP-30 emotional well-being, z = -1.622, P > .05; and EHP-30 social support, z = 0.265, P > .05). However, a significant difference was found on the scale relating to self-image (z = -2.069, P < .05). No significant differences were found between times 1 and 2 on any of the modular scales, and all scales achieved high intraclass correlations exceeding 0.80,15 indicating high test-retest reliability (Table 7).
|
|
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two disease-specific questionnaires have been developed for endometriosis,3,6 but the validity of these instruments is questionable given that most of the items were derived by clinicians and/or were scales taken from generic health-status questionnaires. The EHP-30 is a new, reliable, and valid patient-generated instrument to measure the health-related quality of life of women with endometriosis. The internal reliability of the questionnaire was high, with all the scales exceeding the accepted value of 0.70. It has content validity because the items on the questionnaire were generated from the concerns of the patients rather than relying on the existing literature or clinical assessments of what is important to patients with this condition.
The dimensions produced on the questionnaire reflect the many areas of well-being and functioning that are affected by endometriosis. The generation of the pain scale is consistent with other evidence that pelvic pain has a significant impact on the quality of life of women with endometriosis6 and that chronic pain is one of the most significant aspects of illness. However, other important dimensions, apart from pain, are specific to the experiences of women with endometriosis (ie, feelings of control and powerlessness, lack of social support, self-image, impaired sexual functioning, and worries about infertility). These dimensions have previously been overlooked in the measurement of the health status of women with endometriosis. The EHP-30 has construct validity, because the scales correlated significantly with similar scales of the Short Form-36 as hypothesized and a good negative association was found. The test-retest reliability of the questionnaire was also found to be high, as all the scales achieved high intraclass correlations exceeding 0.8.
The next stage is to evaluate the responsiveness or the "sensitivity to change" of the questionnaire. This is the questionnaire’s ability to measure the effect of treatment on the patient’s health-related quality of life.7 Developing a questionnaire that is able to measure the effect of therapies on health status will be invaluable in endometriosis. The treatment options are limited. New medical therapies and advances in conservative surgery (ie, the use of laser laparoscopy as a treatment for endometriosis) have increased patients’ options in recent years, but the relief that these treatments provide is, at best, short term. It is still only with radical surgery (ie, total abdominal hysterectomy and bilateral salpingo-oophorectomy) that patients can expect a cure, although the symptoms can persist even after this operation.
It is important therefore to evaluate from the patients’ perspective the ways in which these treatments are affecting their health-related quality of life. A patient-generated questionnaire that is "sensitive to change" could be used to assess how these interventions affect women’s daily lives and well-being. Its application in various health care settings would provide new and valuable information not provided by the traditional clinical measures of outcome.
|
Footnotes |
|---|
Received October 17, 2000. Received in revised form March 28, 2001. Accepted April 5, 2001.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Burry KA. Nafarelin in the management of endometriosis: Quality of life assessment. Am J Obstet Gynecol 1992;166: 735–9.[Medline]
3. Bodner C, Garratt AM, Ratcliffe J, MacDonald LM, Penney GC. Measuring health-related quality of life outcomes in women with endometriosis: Results of the gynaecology audit project in Scotland. Health Bull (Edinb) 1997;55: 109–17.
4. Garry R, Clayton R, Hawe J. The effect of endometriosis and its radical laparoscopic excision on quality of life indicators. Br J Obstet Gynaecol 2000;107:44–54.
5. Miller JD. Quantification of endometriosis-associated pain and quality of life during the stimulatory phase of gonadotropin-releasing hormone agonist therapy: A double-blind, randomized, placebo-controlled trial. Am J Obstet Gynecol 2000;182:1483–8.[Medline]
6. Colwell H, Mathias SD, Pasta DJ, Henning JM, Steege JF. A health-related quality of life instrument for symptomatic patients with endometriosis: A validation study. Am J Obstet Gynecol 1998;179:47–55.[Medline]
7. Streiner DL, Norman G. Health measurement scales: A practical guide to their development and use. 2nd ed. Oxford: Oxford University Press, 1995.
8. Patrick DL, Deyo RA. Generic and disease-specific measures in assessing health status and quality of life. Med Care 1989;27:S217–32.[Medline]
9. Fitzpatrick R, Davey C, Buxton M, Jones DR. Patient-assessed outcome measures. In: Black N, Brazier J, Fitzpatrick R, Reeves B, eds. Health services research methods. London: BMJ Publications, 1998:13–22.
10. Jenkinson C, Fitzpatrick R, Brennan C, Bromberg M, Swash M. Development and validation of a short measure of health status for individuals with amyotrophic lateral sclerosis/motor neurone disease: The ALSAQ-40. J Neurol 1999;246(Suppl 3)III:16–21.[Medline]
11. Peto V, Jenkinson C, Fitzpatrick R. PDQ-39: A review of the development, validation and application of a Parkinson’s disease quality of life questionnaire and its associated measures. J Neurol 1998;245(Suppl 1):S10–4.
12. Kline P. A psychometrics primer. London: Free Association Books, 2000.
13. Cronbach L. Coefficent alpha and the internal structure of tests. Psychometrika 1951;16:297–334.
14. Nunnally JC. Psychometric theory. New Delhi: Tate McGraw Hill, 1978.
15. Kline P. A handbook of test construction. London: Methuen & Co, 1986.
This article has been cited by other articles:
|
|
 |
|
  G. Jones, C. Jenkinson, N. Taylor, A. Mills, and S. Kennedy Measuring quality of life in women with endometriosis: tests of data quality, score reliability, response rate and scaling assumptions of the Endometriosis Health Profile Questionnaire Hum. Reprod., October 1, 2006; 21(10): 2686 - 2693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G.L. Jones, K. Benes, T.L. Clark, R. Denham, M.G. Holder, T.J. Haynes, N.C. Mulgrew, K.E. Shepherd, V.H. Wilkinson, M. Singh, et al. The Polycystic Ovary Syndrome Health-Related Quality of Life Questionnaire (PCOSQ): a validation Hum. Reprod., February 1, 2004; 19(2): 371 - 377. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
