Development an Development and Validation of the SEC-QOL Questionnaire in Womend Validation of the SEC-QOL Questionnaire in Women

7/25/2019 Development an Development and Validation of the SEC-QOL Questionnaire in Womend Validation of the SEC-QOL Questionnaire in Women

1/8

Development and Validation of the SEC-QOL Questionnaire in WomenUsing Contraceptive MethodsEzequiel Prez-Campos, MD1, Jose Lus Dueas, MD2, Esther de la Viuda, MD3, Mara ngeles Gmez, MD4,Roberto Lertxundi, MD5, Rafael Snchez-Borrego, MD6, Ignaci Canals, PhD7, Rafael Bermejo, MD7, Agns Arbat, MD7,Xavier Badia, PhD8, Nria Perulero, BSc8, Luis Ignacio Lete, MD9,*1Department of Obstetrics and Gynecology, Hospital General de Requena, Requena, Valencia, Spain; 2Department of Obstetrics and Gynecology, Hospital UniversitarioVirgen Macarena, Sevilla, Spain; 3Department of Obstetrics and Gynecology, Hospital Universitario de Guadalajara, Guadalajara, Spain; 4Centro de Salud Sexual yReproductiva Alicante 3, Alicante, Spain; 5Clnica Euskalduna, Bilbao, Spain; 6Clnica Diatros. Barcelona, Spain; 7Medical Department, Bayer Schering Pharma,Barcelona, Spain; 8IMS Health, Barcelona, Spain; 9Department of Obstetrics and Gynecology, Hospital Santiago Apstol, Vitoria-Gasteiz, Spain

A B S T R A C T

Objectives: Develop and validate a Spanish society of contraceptionquality-of-life (SEC-QOL) questionnaire to assess the impact of contra-ceptive methods on the health-related quality of life (HRQOL) ofwomen. Methods: SEC-QOL was developed following a standardizedprocedure including review of the literature, interviews with contra-ception users, and the administration of a pilot questionnaire to 187women.SEC-QOLconsists of 19 items andincludes fivedimensions. Tovalidate the questionnaire, a multicenter, observational, prospectivestudy was conducted in Spain. The following three study groups weredefined: group A (n 129) comprised women using effective contra-ceptive methods; group B (n 251), comprised women about to startusing an effective method; and group C (n 73) comprised womenusing no or poorly effective contraception. All women attended base-

line and final visits (4

1 months). Participants completed the SEC-QOL, psychological well-being index, EuroQol five-dimensional ques-tionnaire, and perceived health state questionnaires. Results: Atbaseline, women from group A had a better HRQOL in all SEC-QOL

dimensions, except for breast symptoms. Heavier menstrual bleeding,more androgenic and breast symptoms, menstrual pain, and not usinghormonal contraceptive methods were associated with lower HRQOL.SEC-QOL scores showed moderate correlations to psychologicalwell-being index and slightly lower correlation to EuroQol five-di-mensional questionnaire scores. At follow-up, HRQOL had improvedin all groups; most markedly in group B, which obtained an averageeffect size of 0.59. The minimum important difference was estab-lished as a 3.4-point change in the global SEC-QOL score. SEC-QOLobtained a Cronbachs of 0.88 and an intraclass correlation coeffi-cient of 0.82. Conclusions: SEC-QOL is a valid, reliable,and sensitivetochange questionnaire for use in daily clinical practice and future re-search projects on contraception.Keywords: Contraception, Contraceptivemethods, Quality of life, Ques-

tionnaires, Validation.

Copyright 2011, International Society for Pharmacoeconomics andOutcomes Research (ISPOR). Published by Elsevier Inc.

Introduction

The choice of a contraceptive method is mainly influenced by fac-tors related to personal experience acquired from previously usedmethods, includingthe degree of satisfactionwith themethod anditsinfluenceon quality of life in general andsexuality in particular[13].Other elements involved include the effectiveness and easeof useof the method, itspotentialeffects on skin, and itsinfluence

on symptoms associated with the menstrual cycle[4,5](fluid re-tention, breast tension, abdominal swelling, excessive bleeding,and catamenial headache). Effects associated with the menstrualcycle have an influence on health-related quality of life (HRQOL),particularly in women who use hormonal methods[68]that, be-cause of their mechanism of action, significantly decrease mostsymptoms related to the menstrual cycle (premenstrual symp-toms, dysmenorrhea, and amount and duration of bleeding)[9].

To date, the few studies assessing the impact of use of the differ-ent contraceptive methods on HRQOL haveusedgeneric assessment

tools[10],such as instruments specific to psychiatry (quality of lifeenjoymentand satisfaction questionnaire) [8,11,12] or fordirectly as-sessing the impact of menstrual cycle (Moos menstrual distressquestionnaire) [13]. None of these is a specific tool for assessingHRQOL in women of childbearing age using contraceptive methods.The objective of this project was to design and validate a question-naire, the Spanish society of contraception quality-of-life (SEC-QOL)questionnaire, which would allow for specifically assessing the im-pact of contraceptive methods on the HRQOL of women and would

help health-care professionals in contraceptive counseling.

Methods

Questionnaire development

The SEC-QOL questionnaire was developed following a standard-izedprocedure. A comprehensivereview of the scientific literaturewasfirst conducted to identify the questionnaire contents with an

*Addess correspondence to: LuisIgnacio Lete, Department of Obstetrics and Gynecology, HospitalSantiagoApstol, E-01004 Vitoria-Gasteiz.Spain.E-mail:[email protected].

1098-3015/$36.00 see front matter Copyright 2011, International Society for Pharmacoeconomics and Outcomes Research (ISPOR).

Published by Elsevier Inc.doi:10.1016/j.jval.2011.08.1729

V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9

A v a i l a b l e o n l i n e a t w w w . s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w ww . e l s e v i e r . c o m / l o c a t e / j v a l
mailto:[email protected]:[email protected]://www.elsevier.com/locate/jvalhttp://dx.doi.org/10.1016/j.jval.2011.08.1729http://dx.doi.org/10.1016/j.jval.2011.08.1729http://www.elsevier.com/locate/jvalmailto:[email protected]


2/8

expert group. In addition, 14 interviews were also conducted withwomen of childbearing age (aged 1848 years) from different geo-graphic areas of Spain who were using hormonal and nonhor-monal contraceptive methods. The most relevant statements ob-tained from these interviews were used to prepare thequestionnaire, which was subsequently administered to anothergroup, this time consisting of 187 women. A preliminary analysis

of the measurement properties of the questionnaire (internal con-sistency and feasibility) and a Rasch analysis to reduce the num-ber of items were performed using the responses collected[14].The pilot test allowed the number of questionnaire items to bereduced from 41 to 19 and gave a preliminary indication of thegood measurement properties of SEC-QOL.

The final self-administered questionnaire consists of 19 itemsthat allowsfor assessment of the following fivedimensions:social(5 items), menstrual symptoms (4 items), breast symptoms (3items), psychological (4 items), and sexual (3 items) (Appendixfound in Supplemental Materials at 10.1016/j.jval.2011.08.1729).Each item allows for five Likert-like response choices (from al-ways to never or from totally agree to totally disagree de-pending on the type of statement). Scoresfor the overall question-

naire and for each of its dimensions are obtained by adding theresponses to the corresponding items, with subsequent standard-ization to a scale ranging from 0 (worst HRQOL) to 100 (bestHRQOL). Standardization is obtained as follows: (actual score minimum score)/(maximum score minimum score)*100.

Questionnaire validation

To validate the questionnaire, a multicenter, observational, pro-spectivestudywas conductedin Spain from November 2008 to July2009. Three study groups were defined based on the contraceptivemethod used. The first group (group A) included women usingeffective contraceptive methods such as hormonal contracep-tives, consistent use of barrier contraceptives, intrauterine de-

vices, implants and injectables, male and female sterilization, anddual methods (e.g., combination of a hormonal method plus abarrier method). The main requirement was that this group ofwomen continuedusingsuch methods forthe following4 months.Thesecond group (group B)includedwomenwho were notusingany contraceptive method or were using a poorly effective contra-ceptive method (chemical contraceptives, inconsistent condomuse, or natural methods) but who were to start using an effectivemethod from the study recruitment visit and who would use theeffective method for the following 4 months. Finally, the thirdgroup (group C) comprised women who used no or poorly effec-tive contraceptive methods and who, despite contraceptive coun-seling by the physician, decided that they would continue to use apoorly effective contraception or no contraception for the next 4

months. All women with diseases that could directly influencetheir daily life, women who had received a hysterectomy, womenunder study for infertility, women who had given birth within theprevious 6 months, women who were or wanted to become preg-nant, and women participating in a clinical trial at the time ofstudy start were excluded from the study.

The women in Group B who did not continue with the contra-ceptive method of their group at the time of their final visit werereassigned to group C. In other words, the women in group B whodid not use an efficient contraceptive method at the time of theirsecond visit were analyzed as patients of group C. The womenincluded in the group A (effective contraceptive methods) whochanged or didnt continue using an effective contraceptivemethod, were eliminated from the analysis.

The study was approved by the ethics committee of Hospital

Universitario de Guadalajara (Spain).

Sample size

Thesample size required to assessthe measurement properties ofthe SEC-QOL questionnaire was calculated for each study group.Women in group A were included to allow assessment of the testretest reliability of the questionnaire and were observed to obtainan intraclass correlation coefficient of 0.7 or higher, assuming aminimum ratio of 0.4. Assuming a significance level of 0.5, a sta-tistical power of 0.80, and that 15% of patients would be lost tofollow-up or were not evaluable, a sample size of 126 women wasrequired.

Women in group B were included to allow assessment of sensi-tivity to change of the SEC-QOL. The sample size was calculated todetect changes in score of 0.2 standard deviations (small effect size[15])with a significance level of 0.05 and a statistical power of 0.80.Assuming15%of patients wouldbe lost tofollow-up ornot evaluable,the sample size calculated for this group was 231 women.

Finally, women in group C (or control group) were included toallow detection of differences between the other two groups ofwomen using contraceptive methods (groups A and B) in SEC-QOLscores. Forthis,a ratioof onecontrol forevery three womenfrom theother two groups was estimated. A sample size of 77 women wasrequired to detect differences of 0.4 standard deviations or higherwith a significance level of 0.05 and a statistical power of 0.80.

Study measurements

All women recruited into the study (Groups A, B, and C) attendedtwo visits, a baseline visit and a final visit at 4 months (1 month).At the baseline visit, the following were recorded: sociodemo-graphic data (age, place of birth, educational level, and employ-ment status); reason for consultation (regular check-up, contra-ceptive counseling, request for change in contraceptive method);current contraceptive method used and method to be used fromthat visit;menstrual interval (lessthan 21 days, 21 to 35 days, morethan 35 days, and other); bleeding intensity (heavy, normal,scarce, or very scarce); presence or absence of pain related to men-

ses (recorded using a 010 visual analog scale [VAS]); presence ofandrogenic symptoms (acne, alopecia, oily skin, seborrheic der-matitis, hirsutism); presence of premenstrual breast symptoms(pain, increased sensitivity, enlargement, and breast discharge);obstetric history by recording the obstetric formula; and personalhistory of concomitant chronic diseases and life events (change ofpartner or relationship status, occupational changes, changes infamily environment, housing and economic problems). At the fi-nal visit, the same variables were recorded, except for sociodemo-graphic characteristics, reason for consultation, and reason forusing contraceptives.

At both visits, study participants completed the specific SEC-QOL questionnaire, the psychological well-being index (PWI)[16],and the EuroQol five-dimensional (EQ-5D) questionnaire [17]; theywere asked about their perceived health state using a Likert scale

with five response options (from very good to very poor). Allquestionnaires were given to participants for completion beforethe interview with the physician.

Statistical method

A descriptive, comparative analysis of the sociodemographic andclinical characteristics of group A, B, and C study participants wasperformed. For group comparison, an analysis of variance(ANOVA) test with theBonferroni correction wasused formultiplecomparisons of continuous variables, and a chi-square test wasused for categorical variables. Changes in clinical variables andsymptoms from baseline to the final visit were also compared inthe different study groups using the same statistical tests.

The feasibility of SEC-QOL wasassessedfrom the proportion ofmissing responses for each item and the proportion of women

893V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


3/8

with valid responsesto allquestionnaire items. Mean time used tocomplete the questionnaire was also analyzed in groups A and B.

Validity, reliability, and sensitivity to change were assessed asmeasurement properties of SEC-QOL. Discriminant validity wasfirst assessed by comparing the questionnaire scores of the differ-ent study groups, using an ANOVA test. In order to identify anddescribe which variables were related to HRQOL, a bivariate anal-

ysis wasfirst performed, followed by a linearregression, includingwomen from groups A and B and using as dependent variable theglobalSEC-QOL score at thesecond visit (because this wasthe visitforwhich more responsesto the questionnaire from women usingeffective contraceptives were available). All sociodemographicand clinical variables, negative vital events, and the contraceptivemethod used were analyzed as independent variables.

To assess convergent validity, the relationship between thescores in the SEC-QOL and scores obtained in the PWI and VAS ofthe EQ-5D was analyzed using Pearsons correlation coefficient.Women from groupsA andB were included in this analysis.Scoresobtained in the questionnaire were analyzed based on the pres-ence of problems in the items of the descriptive system of theEQ-5D questionnaire using a Studentsttest.

To assess longitudinal validity, changes seen in scores in the

SEC-QOL questionnaire from the baseline to the final visit werecompared to the changes seen in clinical variables (menstrual in-terval, menstrualpain, bleeding intensity, breast symptoms, num-ber of symptoms, and adverse event occurrence) during the sametime period. For this, a Students t test, an ANOVA test, and thePearsons correlation coefficient were used depending on thetype of variable analyzed. To assess sensitivity to change, the ef-fect size (defined as the difference between themean baseline and

final scores, divided by the standard deviation obtained at base-line) obtained in SEC-QOL was calculated. The minimal importantdifference of the questionnaire (defined as the smallest differencein scores of a questionnaire perceived by the patient as beneficial)was estimated as the difference seen by women who stated thattheir health status had slightly improved at 4 months of studystart.

Finally, the internal consistency and testretest reliability ofthe specific questionnaire were assessed. Internal consistencywascalculated using a Cronbachs statistic for all women partic-ipating in the study[18] (for both the overall questionnaire anddimension scores). Testretest reliability was assessed in group Awomen using intraclass correlation coefficient.

SPSS for Windows version 15.0 (SPSS, University of Chicago,Chicago, IL) was used for statistical data analysis. A significancelevel of 0.05 was considered for all group comparisons.

Results

Sample description

A total of 453 women were included. Of these, 129 (28.4%) usedeffective methods and continued to use them after the baselinevisit (group A), 251 (55.4%) did not use effective contraceptivemethods and started to use them from the first study visit (groupB), and 73 (16.1%) did not use effective contraceptive methods anddid not use them throughout the study (group C).Table 1providesthe sociodemographic characteristics of study participants.Women in group C had a slightly higher mean age (33.6 years old)

Table 1 Sociodemographic characteristics of women recruited into the study.

Group A Group B Group C

Age (years), mean (SD)* 29.3 (7.7) 28 (7.4) 33.6 (9.1)

Place of birth, N (%)Spain 115 (89.1%) 207 (82.5%) 63 (86.3%)Eastern Europe 3 (2.3%) 7 (2.8%) 6 (8.2%)Latin America 6 (4.7%) 34 (13.5%) 4 (5.5%)North Africa 1 (0.8%) 2 (0.8%) 0Other 4 (3.1%) 1 (0.4%) 0Total 129 (100%) 251 (100%) 73 (100%)

Educational level, N (%)No education 0 1 (0.4%) 1 (1.4%)Primary education 13 (10.1%) 58 (23.1%) 14 (19.4%)Secondary education 52 (40.3%) 109 (43.4%) 31 (43.1%)University studies 64 (49.6%) 83 (33.1%) 26 (36.1%)Total 129 (100%) 251 (100%) 72 (100%)

Marital status, N (%)Married 43 (33.3%) 85 (34.0%) 34 (47.2%)

Couple living together but not married 27 (20.9%) 46 (18.4%) 11 (15.3%)Stable relationship but not living together 49 (38.0%) 95 (38.0%) 12 (16.7%)No stable relationship 6 (4.7%) 16 (6.4%) 9 (12.5%)Separated/divorced 2 (1.6%) 4 (1.6%) 5 (6.9%)Widow 0 1 (0.4%) 0Unmarried couple 1 (0.8%) 1 (0.4%) 1 (1.4%)No answer 1 (0.8%) 2 (0.8%) 0Total 129 (100%) 250 (100%) 72 (100%)

Employment status, N (%)Working 83 (66.4%) 133 (57.6%) 64 (79.0%)Unemployed 12 (9.6%) 47 (20.3%) 6 (7.4%)Student 28 (22.4%) 50 (21.6%) 11 (13.6%)Other 2 (1.6%) 1 (0.4%) 0Total 129 (100%) 251 (100%) 71 (100%)

*P

0.001;

P

0.005.

894 V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


4/8

than women inthe two other groups(29.3 years old in group A and28.0 years old in group B) (P 0.001). Statistically significant dif-ferences were also seen in educational level; 89.9% of women ingroup A stated that they had secondary or university education,compared with 76.5% and 79.2% in groups B and C, respectively(P 0.005). There were no differences in marital or occupationalstatus.

In the final visit, two of the participants in group A did notaccomplish the conditions of their group; i.e., they used inefficientcontraceptive methods instead of using efficient contraceptivemethods. For this reason those women were excluded from theanalysis population. Final analysis of group A revealed that 127women completed visit 1 and 108women completed visit 2. Of the251 women in group B, 197 completed the second visit. Of those197, 10 did not change their contraception methods (they did notaccomplish the inclusion criteria of group B). Those women ac-complished the inclusion criteria of group C, and finally thosewomen were included in this group. Therefore, for group B weconsidered 241 women in visit 1 and 187 women in visit 2. Thismeans that, for group C, 83 women were considered for the basalvisit (73 10) and 8 of them did not complete the second visit (75women on visit 2).

The main reason for consultation was a regular healthcheck-up in 78.7% and 77.1% of women in groups A and C, respec-tively, whereas the most common reason (56.4%) in group B wascontraceptive counseling. Among women who did not use anyeffective contraceptive method at baseline (women from groups Band C), the most common method was inconsistent condom use(78.8% and 62.7%, respectively), followed by coitus interruptus(18.7% and 24.1%). Among women usingan effective contraceptivemethod at the baseline visit (group A), 42.5% stated that they usedthecontraceptive pill and24.4%used condoms consistently(Table2). With respect to the mean (SD) number of years women hadbeen using the same contraceptive method, the longest periodwas found for group C women (6.4 [6.9], compared with 4.0 [3.9]and 3.8 [4.5] for women from groups A and B, respectively; P 0.001). Of women from group B who switched from a poorly effec-

tive to an effective contraceptive method, 38.6% were planning tochange to the contraceptive pill. Besides contraception, regulationof the menstrual cycle was reported as another reason for startingcontraceptive use by 23.6% of group A women and by 12.0% ofgroup B women.

At thebaselinevisit, more than 84%of women in allgroups hadnormal menstrual cycles (menstrual interval ranging from 21 and35 days), with no significant differences between the differentstudy groups. By contrast, statistically significantdifferenceswerefound in bleeding intensity, which was greater among women

from groups B and C than those from group A (P 0.001). For thissame variable, a statistically significant decrease was seen fromthe baseline to the final visit in group B women only (P 0.001).With respect to menstrual pain, group A women had less painthan those from the two other groups (P 0.001). However, a sta-tistically significant decrease from the baseline to the final visitwas only seen in group B women (P 0.001). The most common

androgenic symptoms were acne, oily skin, andgreasy hair, butnostatistically significant differences were seen between the groups.The number of androgenic symptoms decreased in group Bwomen from the baseline to the final visit (P 0.001). With regardto breast symptoms, themost commonwas an increasedpremen-strualbreast tenderness, occurringin 44.4%of group A women andin 56.5% and 67.5% of women from groups B and C, respectively.Differences were seen between the groupsin thenumber of breastsymptoms. Women from group B were theonly ones whoreporteda statistically significant decrease in the number of breast symp-toms from the baseline to the final visit ( P 0.002).

Validation of SEC-QOL

FeasibilityA total of 96.4% women answered 100% of the items of SEC-QOL.The missing response rate was less than 2% in all questionnaireitems. Mean administration time was 6.1 minutes.

ValidationFigure 1shows the baseline scores obtained in the SEC-QOL ques-tionnaire by the different study groups. Statistically significantdifferences were seen between group A women (using effectivecontraceptive methods) and women from groups B and C (not us-ing effective contraceptive methods until the time of visit) in alldimensions (P 0.05), except for breast symptoms. Women fromgroup A had a better HRQOL. Statistical significant differences inHRQOL between study groups were confirmed (P 0.01) using aregression model adjusting by age and marital status.

When variables related to HRQOL were assessed by jointly an-alyzing women from groups A and B, a relationship was seen withmenstrual bleeding and pain, but not with duration of menstrualinterval. Women with a greater menstrual bleeding intensity alsorevealed reduced HRQOL according to the global SEC-QOL score(P 0.01) as well as the scores for the menstrual symptom ( P 0.05) and psychological (p0.01) dimensions. Women with moreseveremenstrual pain also indicated reduced HRQOL according toall SEC-QOL scores (global and by dimensions) (P 0.02 for thesexual dimension andP 0.01 for all other scores).

Table 2 Contraceptive methods used at the baseline visit by study group.

Group Bn (%)

Group Cn (%)

Group An (%)

None 23 (9.5%) 14 (16.9%) Pill 54 (42.5%)Spermicide 0 1 (1.2%) Gestagen-only pill 1 (0.8%)Ovules 1 (0.4%) 0 Patch 1 (0.8%)Creams 0 1 (1.2%) Ring 16 (12.6%)Gels 0 0 Implant 1 (0.8%)Inconsistent condom 190 (78.8%) 52 (62.7%) Levonorgestrel IUD 12 (9.4%)Coitus interruptus 45 (18.7%) 20 (24.1%) Injectable 1 (0.8%)Ogino 4 (1.7%) 2 (2.4%) Condom 31 (24.4%)Temperature 0 1 (1.2%) IUD 6 (4.7%)Symptothermal 0 1 (1.2%) Vasectomy 3 (2.4%)Total 241 (100%) 83 (100%) Hormonal condom 3 (2.4%)

Total 127 (100%)

IUD, intrauterine device.

895V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


5/8

The number of breast and androgenic symptoms experiencedby study participants wasfound to have an impacton HRQOL.Thenumber of breast symptoms experienced indicated a Pearsonscorrelation coefficient of 0.64 (P 0.01) with the breast symptomdimension and a coefficient ranging from 0.13 to 0.37 (P 0.05)with allother dimensions, except for the sexual dimension, whichhad a very low Pearsons correlation coefficient of 0.01 ( P 0.94).The number of androgenic symptoms demonstrated a statisticallysignificant but small (0.133) Pearsons correlation coefficient withthe global SEC-QOL score.

A bivariate analysis revealed that women with a poorerHRQOLincluded those with greater menstrual bleeding intensity, moreandrogenic and breast symptoms, menstrual pain, and patients

who did not use hormonal contraceptive methods. Subsequently,in the multivariate regression model, the variables found to besignificant included breast symptoms (the greater their presence,the lower HRQOL), menstrual pain intensity (the greater the pain,the poorer HRQOL), and use of hormonal methods (better HRQOLamong users of these methods) (Table 3).

SEC-QOL scores indicated moderate correlations to PWIscores. The highest correlation coefficients were found in thepsychological dimension of SEC-QOL, followed by global score.The social and menstrual symptom dimensions of the SEC-QOLhad a greater correlation to the vitality and general health di-mensions of PWI than to all other questionnaire dimensions(Table 4). The correlations observed in the EQ-5D VAS are

59.2

65.2 63.9

55.0

72.2

53.8

46.6

53.7

46.047.7

60.1

38.8

49.8

54.9 54.9

43.9

60.0

42.9

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

Total Social D. Menstrual

Symptoms

Breast

Symptoms

Psychological

D.

Sexual D.

SEC-QOLSco

re


A vs B

A vs C A vs B

A vs C*

A vs B

A vs C*

B vs C*

A vs C*A vs B

A vs CA vs B

A vs C

Fig. 1 Mean baseline total and dimension scores by study group.

Table 3 Coefficients obtained in the bivariate and multivariate regression models to analyze variables related to theglobal SEC-QOL score.

Bivariate regressions Multivariate model

Independent variables at final visit Beta Pvalue Beta Pvalue

Breast symptoms 0.300 0.001 0.210 0.001Menstrual pain 0.375 0.001 0.297 0.001Use hormonal method 0.186 0.001 0.130 0.035Bleeding intensity 0.188 0.001 0.032 0.614Androgenic symptoms 0.132 0.031 0.077 0.185Use levonor. IUD 0.150 0.011Use condoms 0.148 0.012Use IUD 0.121 0.040Use pill 0.083 0.159Use ring 0.048 0.414Use patch 0.111 0.060Age 0.098 0.097Menstrual interval 0.001 0.984Some bad event 0.091 0.121

IUD, intrauterine device.

896 V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


6/8

slightly lower and reach the maximum coefficient with theglobal SEC-QOL score. When SEC-QOL scores were analyzed ac-cording to the level of problems in each of the EQ-5D dimen-sions, the greatest differences were seen in relation to pain/discomfort, anxiety/depression, and daily activities (Table 4).The relationship between the question about the perceivedhealth state and SEC-QOL was statistically significant for alldimensions (P 0.01) except for the sexual dimension.

As presented in Figure 2,HRQOL tended to improve in all

study groups from the baseline to the final visit, but this im-provement was more marked in women from group B. Onlygroup A women had statistically significant improvements in

the overall SEC-QOL score (P 0.05) and in the psychologicaldimension (P 0.01), whereas group B women had improve-ments in global score (P 0.01) as well as the social (P 0.01),psychological (P 0.01), and sexual (P 0.01) dimensions.Group B women had a HRQOL similar to group C women atbaseline, but at the final visit their HRQOL was more similar tothat of group A women (P 0.01).

Sensitivity to changeThe effect size (ES) obtained forall SEC-QOL scoreswere greater ingroup B women (0.14 to 0.59) compared with those in group A (0.06

Table 4 Correlations between SEC-QOL scores and the PWI questionnaire.

Global score Socialdimension

Menstrualsymptoms

Breastsymptoms

Psychologicaldimension

Sexualdimension

PWI (Pearsonsr)Total 0.404* 0.350* 0.320* 0.128 0.433* 0.167*

Anxiety 0.347* 0.273* 0.302* 0.115

0.390* 0.138*Depression 0.283* 0.256* 0.196* 0.089 0.334* 0.114

Positive mood 0.270* 0.253* 0.180* 0.051 0.316* 0.174*Vitality 0.412* 0.387* 0.326* 0.136 0.374* 0.162*Self-control 0.278* 0.224* 0.215* 0.062 0.326* 0.127

General health 0.405* 0.359* 0.345* 0.168* 0.399* 0.127

EQ-5D VAS (Pearsonsr)VAS 0.239* 0.224* 0.206* 0.106 0.218* 0.079

EQ-5D descriptive system (Pvalue)Mobility 0.169 0.633 0.251 0.909 0.068 0.01Self-care 0.153 0.370 0.538 0.937 0.211 0.347Daily activities 0.01 0.05 0.304 0.173 0.01 0.631Pain/discomfort 0.01 0.01 0.01 0.062 0.01 0.191Anxiety/depression 0.01 0.01 0.123 0.144 0.01 0.101

Health state (Pvalue)

Health state (010) 0.01 0.01 0.01 0.01 0.01 0.065EQ-5D, EuroQol questionnaire; PWI, psychological well-being index; SEC-QOL, Spanish society of contraception quality-of-life questionnaire;VAS, visual analogue scale.*P 0.01; P 0.05.

30

40

50

60

70

80

90

baseline

final

baseline

final

baseline

final

baseline

final

baseline

final

baseline

final

Total Social D. Menstrual

Symptoms

Breast

Symptoms

Psychological

D.

Sexual D.

SEC-QOLScore


Group A*

Group B

Group C

BpuorGBpuorGBpuorG

Group A

Group B

Group Cns

Fig. 2 Changes in scores obtained in each of the dimensions of the specific questionnaire (SEC-QOL) by study group.*P< 0.05; P< 0.01.

897V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


7/8

to 0.24). In the overall score, group B women obtained an ES of 0.59(large ES), compared with an ES of 0.22 for group A women (smallES) (Table 5).

The minimum important difference in this study was estab-lishedas a 3.4-pointchange in theglobal SEC-QOL score.Minimumimportant difference is defined as the mean change in question-naire scores perceived by patients as an improvement in HRQOL.

ReliabilityThe SEC-QOL questionnaire demonstrated a good internal consis-tency,with an overall Cronbachs of 0.88 and a value higherthan0.70 in all dimensions, except for the sexual dimension (0.55). Theoverall intraclass correlation coefficient was 0.82, and was higherthan 0.70 in all dimensions, thus indicating the good testretestreliability of the questionnaire.

Discussion

Our results show the high feasibility, validity, sensitivity tochange, and reliability of the SEC-QOL as a specific questionnaire

to assess HRQOL in women of childbearing age who are usingcontraceptive methods. The SEC-QOL consists of only 19 itemsand may be administered in a short time. It is thus considered tobe suitable for use in standard clinical practice.

Various studies on HRQOL and contraception have demon-strated that symptoms related to the menstrual cycle (fluid reten-tion, breast tension, abdominal swelling, skin changes, heavybleeding, or migraine associated to menstrual cycle), as well assymptoms such as depression, irritability, or mood changes arerelated to a poorer HRQOL[6,19,20],which especially affects thephysical dimension and mood in women. The impact of premen-strual symptoms on the daily life of women will depend on sever-ity of such symptoms [21]. Because of the high prevalence of thosesymptoms in women of childbearing age (estimated to representmorethan70% in the United States [22]), it is important to individ-

ually assess not only the number of symptoms, but the impact ofregular occurrence of these symptoms on the life of women. Theresults achieved using the SEC-QOL questionnaire, developed us-ing a methodology taking into account, in addition to publishedstudies andexpert opinions,the opinion of women of childbearingage using and not using contraceptive methods, reveal that previ-ous or ongoing premenstrual symptoms have the greatest impacton the daily life of women. By contrast, aspects related to theefficacy of contraceptive methods do not appear to have a verysignificant impact on HRQOL because the questionnaire only in-cludes one question directly related to this aspect.

With regard to premenstrual symptoms, this study found thatpremenstrual breast tension and intensity of menstrual bleedingare symptoms with a great impact on the life of women who ex-perience them. Women already using some effective contracep-

tive method at the baseline visit (group A) had better scores in themenstrual and breast symptom dimensions than women whowere to start an effective contraceptive method (group B). In addi-tion, women who start a hormonal contraceptive method havealso been shown to experience a significant improvement insymptoms related to the menstrual cycle during the study, animprovement also reflected in an increase in SEC-QOL scores. Therelationship found between a greater bleeding intensity, a greaternumber of androgenic and breast symptoms, a higher severity ofmenstrual pain, and the use of nonhormonal contraceptive meth-ods and a poorer quality of life agree with other studies relatinguse of hormonal contraceptives to a decrease in premenstrualsymptoms[3,23,24].

A study intended to assess the impact of oral contraceptivemethods on quality of life (as assessed by the World Health Orga-nization quality of life [WHOQOL]) in a Japanese population con- T

able5EffectsizeoftheSEC-QOL.

Socialdimension/dailyactivities

Menstrualsymptomdimension

Breastsymptomdime

nsion

N

baseline/final

Score

baseline/final

Effectsize

N

baseline

/final

Score

baseline/final

Effectsize

N

baseline/final

Score

baseline/fina

l

Effectsize

G

roupA

126/108

6

5.2

/66.7

0.0

6

126/1

06

63.9

/68.0

0.1

8

126/107

55.0

/57.2

0.0

7

G

roupB

236/187

5

3.7

/60.4

0.2

8

237/1

88

46.0

/61.4

0.5

8

238/187

47.7

/52.2

0.1

4

G

roupA

B

362/295

5

7.7

/62.7

0.2

0

363/2

94

52.2

/63.8

0.4

3

364/294

50.2

/54.0

0.1

2

Psychologicaldimension

Sexualdimension

Totalscore

N

baseline/final

Score

baseline/final

Effectsize

N

baseline

/final

Score

baseline/final

Effectsize

N

baseline/final

Score

baseline/fina

l

Effectsize

G

roupA

126/107

7

2.2

/80.2

0.2

4

126/1

08

53.8

/56.3

0.1

0

126/105

59.2

/63.0

0.2

2

G

roupB

233/186

6

0.1

/74.2

0.4

7

235/1

87

38.8

/51.5

0.5

4

229/184

46.6

/57.0

0.5

9

G

roupA

B

359/293

6

4.3

/76.4

0.3

8

361/2

95

44.0

/53.3

0.3

7

355/289

51.1

/59.1

0.4

3

S

EC-QOL,

Spanishsocietyofcontraceptionq

uality-of-lifequestionnaire.

898 V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9


8/8

cluded that premenstrual symptoms had a great impact on thesocial, physical, and psychological dimensions[11].These resultsalso agree with those seen in theSEC-QOL,suggesting that group Bwomen (with more premenstrual symptoms) have a greater im-pairment in the psychological and social dimensions than group Awomen (with less premenstrual symptoms). However, unlike ge-neric questionnaires, the SEC-QOL includes specific dimensions

related to premenstrual (breast) and menstrual symptoms thatprovide more specific information than the generic HRQOL ques-tionnaires used to date, including the WHOQOL or short form 36health survey. Results of this study indicate that the HRQOL ofwomen of childbearing age who use contraception methods ishighly related to severity of premenstrual symptoms and that thechoice of hormonal contraceptive methods is highly focused onthe minimization of such symptoms; not just obtaining the con-traceptive effect.

Although no change wasmade in the category of contraceptivemethod used by group A women during the study follow-up pe-riod, these women indicated a slight improvement in HRQOL.Such improvement may have been due, first of all, to the partici-pation of women in this study, which resulted in closer medicalmonitoring of women (more visits and longer duration of visits)

and an assessment by the physician of more relevant aspects per-ceptible by patients, such as HRQOL. Second, the fact that about12% of group A women changed from using condom to hormonalcontraceptive methods may have an impact on HRQOL.

Themain limitation of this study is thefollow-up period, whichlasted 4 months, previous contraception studies defined a fol-low-up period between 3 and 6 months. Although 4 months maynot appear long enough to assess changes in HRQOL score, partic-ularlythoserelated to premenstrualsymptomsin group B women,this limitation is minimized by the finding of a high sensitivity tochange (large ES) of the SEC-QOL in this group of women, whichmeans that the questionnaire is able to detect even small changesin the health state of women.

In conclusion, this study permitted the evaluation of the mea-surement properties of the first HRQOL questionnaire specific for

women of childbearing age who are using contraceptive methods.Theresults achieved suggest that theSEC-QOL is valid foruse bothin daily clinical practice and in future research projects to assessHRQOL of women on contraception. Finally, this study also dem-onstrated the significance and impact of premenstrual symptomson HRQOL experienced by women of childbearing age.

Source of financial support: This study was promoted by theSpanish Society of Contraception (SEC, Sociedad Espaola de Con-tracepcin) and funded by Bayer. The authors who received fund-ingwere Ezequiel Prez-Campos, MD,Jose Lus DueasMD, Estherde la Viuda,MD, Mara ngeles Gmez,MD, Roberto LertxundiMD,Rafael Snchez-Borrego, MD, Xavier Badia PhD, Nria PeruleroBSc, and Lus Ignacio Lete, MD.

Supplemental Materials

Supplemental material accompanying this article can be found inthe online version as a hyperlink atdoi:10.1016/j.jval.2011.08.1729or, if a hard copy of article, at www.valueinhealthjournal.com/issues(select volume, issue, and article).

R E F E R E N C E S

[1] Rosenfield JA, Zahorik PM, Saint W, Murphy G. Womens satisfactionwith birth control. J Fam Pract 1993;36:16973.

[2] Oddens BJ. Determinants of contraceptive use among women ofreproductive age in Great Britain and Germany II: psychological

factors. J Biosoc Sci 1997;29:43770.[3] Oddens BJ. Womens satisfaction with birth control: a populationsurvey of physical and psychological effects of oral contraceptives,intrauterine devices, condoms, natural family planning, andsterilization among 1466 women. Contraception 1999;59:27786.

[4] Boschitsch E, Skarabis H, Wuttke W, Heithecker R. The acceptability ofa novel oral contraceptive containing drospirenone and its effect ofwell-being. Eur J Contracept Reprod Health Care 2000;5:34 40.

[5] Bitzer J, Kunz J. Contraception from the viewpoint of women: long-term pill users are especially pleased by the advantages of hormonalcontraception. Comprehensive study of motives at first and refillprescription of the pill and evaluation after 3 months of hormonalcontraception. Contraception 1999;59:28791.

[6] Foidart JM. Added benefits of drospirenone for compliance. Climateric2005;8(Suppl. 3):28 34.

[7] Sulak PJ, Carl J, Gopalakrishnan I, et al. Outcomes of extended oralcontraceptive regimens with a shortened hormone-free interval tomanage breakthrough bleeding. Contraception 2004;70:2817.

[8] Egarter C, Topcuoglu MA, Imhof M, Huber J. Low dose oral

contraceptives and quality of life. Contraception 1999;59:28791.[9] ESHRE Capri Workshop Group. Noncontraceptive health benefits of

combined oral contraception. Hum Reprod Update 2005;11:51325.[10] Li RHW, Lo SST, Teh DKG, et al. Impact of common contraceptive

methods on quality of life and sexual function in Hong Kong Chinesewomen. Contraception 2004;70:474 82.

[11] Matsumoto Y, Yamabe S, Ideta K, Kawabata M. Impact of use ofcombined oral contraceptive pill on the quality of life of Japanesewomen. J Obstet Gynaecol Res 2007;33:52935.

[12] Zhao J, Li Y, Wu Y, et al. Impact of different contraceptive methods onquality of life in rural women of the Jiangsu province in China.Contraception 2009;80:1806.

[13] Legro RS, Pauli JG, Kunselman AR, et al. Effects of continuous versuscyclical oral contraception: a randomized controlled trial. J ClinEndocrinol Metab 2008;93:4209.

[14] Young TA, Yang Y, Brazier JE, Tsuchiya A. The use of Rasch analysis inreducing a large condition-specific instrument for preferencevaluation: the case of moving from AQLQ to AQL-5D. Med Decis

Making 2011;31:195210.[15] Cohen J. Statistical Power Analysis for Behavioral Sciences (2nd ed.).Hillsdale, NJ: Lawrence Erlbaum Associates Inc., 1988.

[16] Badia X, Gutierrez F, Wiklund I, Alonso J. Validity and reliability ofSpanish version of the general well being index. Quality Life Res 1996;5:1018.

[17] Badia X, Roset M, Monserrat S, et al. La versin espaola del EuroQol:descripcin y aplicaciones [in Spanish]. Med Clin 1999;112(Suppl. 1):7985.

[18] Scientific Advisory Committee of the Medical Outcomes Trust.Assessing health status and quality-of-life instruments: attributes andreview criteria. Qual Life Res 2002;11:193205.

[19] Zupi E, Marconi D, Sbracia M, et al. Add back therapy in the treatmentof endometriosis associated pain. Feril Steril 2004;82:13038.

[20] Sulak PJ, Carl J, Gopalakrishnan I, et al. Outcomes of extended oralcontraceptive regimens with a shortened hormone-free interval tomanage breakthrough bleeding. Contraception 2004;70:2817.

[21] Mishell Dr. Premenstrual disorders: epidemiology and disease burden.Am J Manag Care 2005; 11:S4739.

[22] Ginsburg KA, Dinsay R. Premenstrual syndrome. In: Ransom SB, ed.Practical Strategies in Obstetrics and Gynecology. Philadelphia, PA:WB Saunders, 2000.

[23] Clayton AH. Symptoms related to the menstrual cycle: diagnosis,prevalence, and treatment. J Psychiatr Pract 2008;14:1321.

[24] Young EA, Kornstein SG, Harvey AT, et al. Influences of hormone-based contraception on depressive symptoms in premenopausalwomen with major depression. Psychoneuroendocrinology 2007;32:84353.

899V A L U E I N H E A L T H 1 4 ( 2 0 1 1 ) 8 9 2 8 9 9
http://dx.doi.org/10.1016/j.jval.2011.08.1729http://www.valueinhealthjournal.com/issueshttp://www.valueinhealthjournal.com/issueshttp://www.valueinhealthjournal.com/issueshttp://www.valueinhealthjournal.com/issueshttp://www.valueinhealthjournal.com/issueshttp://dx.doi.org/10.1016/j.jval.2011.08.1729

Documents

Development an Development and Validation of the SEC-QOL Questionnaire in Womend Validation of the SEC-QOL Questionnaire in Women