Elsevier

Maturitas

Volume 50, Issue 3, 14 March 2005, Pages 209-221
Maturitas

Further psychometric property development of the Menopause-Specific Quality of Life questionnaire and development of a modified version, MENQOL-Intervention questionnaire

https://doi.org/10.1016/j.maturitas.2004.06.015Get rights and content

Abstract

Objectives: To develop the 1996 MENQOL questionnaire further with advice regarding summary score computation, missing-data management, readability, recall period and assessment of the vasomotor domain reliability and construct validity. To develop a modified version, the MENQOL-Intervention questionnaire, for use where certain treatment side effects could negatively impact the quality of life. Methods: MENQOL-Intervention modifications involved the addition of three items to the physical domain. For both questionnaires, psychometric property assessment was embedded in two randomized controlled trials of menopause interventions. Test–retest reliability and Cronbach’s alpha were computed for all domains as was construct validity of the vasomotor domain for both questionnaires. Results: The vasomotor intraclass correlation coefficient was 0.73 for the MENQOL-Intervention over 1 week and 0.78 for the MENQOL over 1 month. The altered physical domain of the MENQOL-Intervention questionnaire continued to show strong test–retest reliability and Cronbach’s alpha consistent with the MENQOL. The MENQOL-Intervention demonstrated excellent face validity with high construct validity for the vasomotor domain of 0.78–0.80. For both instruments, comparisons of the vasomotor domains to hot flash scores, although statistically significant, were only moderate at 0.56 and 0.49. Conclusions: Both the MENQOL and the MENQOL-Intervention questionnaires show strong psychometric properties. We recommend using the MENQOL-Intervention questionnaire where intervention side effects might negatively impact a woman’s quality of life. For both questionnaires, a summary score can be calculated.

Introduction

The development and initial overview of the psychometric properties of the Menopause-Specific Quality of Life questionnaire (MENQOL) was published in 1996 [1]. Since then, the MENQOL has been used as an outcome measure and has been translated into Finnish, Italian, Polish, Danish, Norwegian, Dutch for the Netherlands and Belgium, French for France, Belgium and Canada, Spanish for Spain, Mexico, Puerto Rico and Argentina, and Portuguese for Brazil. All translations have included linguistic and cultural adaptation issues [2]. A translated version of the MENQOL questionnaire is not appropriate for Cantonese speaking Chinese women in mainland China [3].

This article describes changes made to items, readability, summary score issues, and missing-data management strategies. In addition, we have continued to establish the psychometric properties of the instrument by further work on reliability and construct validity of the vasomotor domain, which had the least robust characteristics during initial development work. The test–retest reliability of 0.37 was low and the original comparison instrument, the somatic subscale of the Neugarten and Kraines Menopause Symptom Checklist [4], was subsequently felt to have been inadequate for vasomotor construct validation.

The Menopause-Specific Quality of Life instrument was developed as a self-administered condition-specific tool to measure health-related quality of life [5] in middle-aged women in the years immediately beyond the onset of menopause, defined as one year with no menses. It was designed to have both discriminative and evaluative characteristics to assess the impact of this life event and to evaluate interventions [6], [7]. The item-generation phase of the questionnaire development, resulting in 106 items, reflected dimensions of physical, emotional, sexual, intellectual, sleep and social functioning, including effects on working life, which might impact on quality of life after a woman is in menopause.

The reduction phase of the questionnaire and establishment of its psychometric properties followed a method championed by Guyatt [8], [9]. In particular, reduction did not use factor analysis, which, as theory suggests, may be an inappropriate method to discern items, which may influence perceived quality of life [10]. An importance score was calculated based on the occurrence and perceived bothersomeness of each item. The mean importance score of each item was used to rank the items for inclusion in the final questionnaire. The 106 items were assigned to dimensions by two clinicians and one social scientist independently with consensus negotiation. Items representing sleep and physical functioning were combined into a physical dimension, whereas vasomotor and sexual items were left separate from the physical dimension. Emotional, intellectual and social functioning was combined into a psychosocial dimension. The final 29-item questionnaire included the highest three items ranked by importance score from each combined dimension, now renamed a domain, followed by the remaining items, ranked together by importance score and assigned according to their dimension of origin to the four domains of the questionnaire. Six items representing a working life dimension were clustered together but dropped from the final questionnaire because the importance scores of the items were markedly lower than the lowest ranking importance score in any domain.

Reliability, responsiveness and validity were assessed and published [1]. Validity assessment included face, content and construct validity. Construct validity was assessed by domain, using comparison questionnaires and strategies, appropriate to the purpose of each domain.

It has been recognized in the literature that consideration should be given to assessing both the positive effects of an intervention and any potential side effects, which could negatively impact the individual’s assessment of perceived health-related quality of life when designing outcome measures to assess treatment [11], [12]. Therefore, the authors developed a modified version, the Menopause-Specific Quality of Life-Intervention questionnaire (MENQOL-Intervention) to ensure that side effects of specific therapies that might adversely affect quality of life were reflected in the outcome measure. The development and psychometric testing of the MENQOL-Intervention questionnaire version are also reported in this paper including test–retest reliability, Cronbach’s alpha, and face validity.

Section snippets

The MENQOL questionnaire format and item changes

The MENQOL was developed on women aged 47–62 years, who were 2–7 years beyond menopause, and it consists of 29 items divided into four domains: vasomotor (three items), psychosocial (seven items), physical (16 items) and sexual (three items). All items follow the same format in which the woman is asked whether she has experienced the item in the previous month. If ‘no’, she goes to the next item; however, if ‘yes’, she indicates how bothered she was by the item on a 7-point Likert scale [13]

Development of the MENQOL-Intervention questionnaire

The modified version, the MENQOL-Intervention questionnaire, includes all MENQOL items plus three items, which might negatively affect quality of life, ‘breast tenderness’ and ‘vaginal bleeding’ with hormone replacement therapy and ‘leg cramps’ with selective estrogen receptor modifiers expanding the physical domain to 19 items. These added items are listed as common side effects in the product reference manual [16], and their importance is verified by clinical experience. The

Use of a summary score

The original MENQOL instrument did not suggest how to form a summary score. Further analysis of the original importance scores calculated in the reduction phase and used to determine, which items entered the MENQOL questionnaire, showed the mean importance scores of the items in each domain to be similar: vasomotor 2.3, psychosocial 1.6, physical 1.7 and sexual 1.7. The women who responded in the reduction phase suggested that they were equally impacted by the cluster of items, which form each

Psychometric property determination for the MENQOL and the MENQOL-Intervention questionnaires

Further analyses of psychometric properties occurred within two separate intervention trials: one trial by Lewis et al. [18] using the MENQOL questionnaire and the other by Gelfand et al. [19] using the MENQOL-Intervention questionnaire. Table 1 summarizes the trial characteristics for the original MENQOL instrument development work by Hilditch et al. [1] along with the two new trials.

The trial by Lewis et al. [18] was a three-arm RCT comparing the effects of soy, flax and wheat (placebo)

Pretest and face validity of the MENQOL-Intervention questionnaire

Nine women provided their assessment of the clarity and ease of understanding of the three additional questions with all finding vaginal bleeding or spotting clear and easy to understand. Eight women rated leg pains or cramps as clear and easy to understand with one woman asking if the question included feet. For the item, breast pain or tenderness, three women suggested that ‘pain’ might be too strong and ‘sensitivity’ might be a better word. In assessing the entire questionnaire’s ability to

Discussion

Modifications to the MENQOL questionnaire include minor rewording and the addition of a summary score. These changes are also used in the MENQOL-Intervention questionnaire with the addition of three new items in the physical domain. Domain internal consistency for the MENQOL-Intervention questionnaire ranged from 0.72 to 0.88 with the physical domain 0.88 similar to the Cronbach’s alpha of 0.87 in the original 16-item MENQOL physical domain by Hilditch et al. and the MENQOL reanalysis of 0.89

Conclusions

The MENQOL-Intervention questionnaire demonstrates good domain internal consistency and test–retest reliability. Since the MENQOL-Intervention questionnaire is completely similar to the MENQOL questionnaire with the exception of the physical domain and the recall period, full validity was not redone. However, we consider that much of the work from the development of the MENQOL questionnaire is transferable to the MENQOL-Intervention questionnaire and vice versa. In particular, the vasomotor

Acknowledgments

The development and testing of the MENQOL-Intervention was supported by Janssen-Ortho, Inc., Canada. E.Y.Y. Wong performed the pretest and face validity.

References (37)

  • J.R. Hilditch et al.

    Experience of menopausal symptoms by Chinese and Canadian women

    Climacteric

    (1999)
  • B. Neugarten et al.

    Menopausal symptoms in women of various ages

    Psychosom Med

    (1965)
  • G.H. Guyatt et al.

    Measuring quality of life in clinical trials: a taxonomy and review

    CMAJ

    (1989)
  • R. Deyo et al.

    Reproducibility and responsiveness of health status measures: statistics and strategies for evaluation

    Control Clin Trials

    (1991)
  • G.H. Guyatt et al.

    Measuring disease-specific quality of life in clinical trials

    CMAJ

    (1986)
  • R. Jaeschke et al.

    Quality of life assessments in clinical trials

    (1990)
  • P.M. Fayers et al.

    Factor analysis, causal indicators and quality of life

    Qual Life Res

    (1997)
  • G.H. Guyatt et al.

    Users’ guides to the medical literature. XII How to use articles about health-related quality of life: evidence-based medicine working group

    JAMA

    (1997)
  • Cited by (0)

    1

    Retired.

    View full text