Sex, drugs, and hypertension: a methodological approach for studying a sensitive subject

Transcription

1 (2001) 13, 31±40 ß 2001 Nature Publishing Group All rights reserved /01 $ Sex, drugs, and hypertension: a methodological approach for studying a sensitive subject LE Duncan 1, C Lewis 1 *, CE Smith 1, P Jenkins 1, M Nichols 1 and TA Pearson 2 1 Research Institute, Bassett Healthcare, One Atwell Road, Cooperstown, NY 13326, USA; and 2 University of Rochester, Department of Community and Preventive Medicine, Rochester, NY, USA Sexual dysfunction is a recognized side effect of hypertension and antihypertensive medications in men, but is not established as a side effect in women, due to the lack of established methodology. An ambulatory medical record-based, case ± control study was designed to study sexual function in treated and untreated hypertensive women and healthy controls. The research was performed in a teaching hospital with satellite clinics in upstate New York. There were 3312 medical records reviewed, 640 premenopausal Caucasian women in heterosexual relationships subjects were eligible for participatic diagnosis of mild hypertension (BP140=90 mmhg and < 160=100 mmhg) for cases; no other signi cant medical history. A total of 241 women agreed to participate, 224 (35%) completed both a self-administered questionnaire and a telephone interview (112 healthy, 112 hypertensive). There was an initial 74% response rate among those eligible to participate, with 35% completing the entire study. Age and average blood pressure were not signi cant between 224 participants and 416 nonparticipants by 2-tailed t-test analysis. Seven composite variables were formed from a 47-item sexual response questionnaire. Initial unadjusted w 2 results reported women with hypertension had more dif culty than did healthy controls achieving lubrication and orgasm. Seven questions, each with the highest correlation to its respective composite variable (by Spearman's correlation), formed an abbreviated questionnaire. Quality of female sexual function was quanti ed in an ambulatory outpatient setting. A method was described to address hypertension, pharmacotherapy, and sexual functioning by employing self-administered questionnaires and telephone interviews. Initial analysis suggested that hypertensive women may have an impaired physiological sexual response. The abbreviated questionnaire generated from questions with the highest correlation to their respective composite variables may be useful in further evaluating this issue. International Journal of Impotence Research (2001) 13, 31±40. Keywords: hypertension; antihypertensive therapy; sexual dysfunction; quality of life; chemicallyinduced psychosexual disorders and dysfunction Introduction The effects of unmedicated hypertension and antihypertensive medication on sexual functioning in women are discussed in the literature but inadequately studied, 1±3 perhaps due to the complex nature of female sexuality and to a lack of tested methodology. A recent literature review 4 reported that few studies during the last two decades examined quality of life and sexual side effects in hypertensive women alone, though sexual dysfunction has been reported as higher in untreated hypertensive males than in normotensive males. 5 Most clinical trials studying the effects of antihypertensive medications on women do not inquire *Correspondence: C Lewis, Research Institute, Bassett Healthcare, One Atwell Road, Cooperstown, NY 13326, USA. Received 28 July 2000; accepted 14 October 2000 in any detail into sexual functioning, with the exception of the Treatment of Mild Hypertension Study (TOMHS). 6 A review in proposed a brief sexual questionnaire, modi ed from other surveys, for use with hypertensive women. 7,8 There is not a standard de nition of normal sexual functioning amongst healthy women in the general population, making it dif cult to compare diminished quality of sexual function in those with this chronic condition. A study of sexual history in general medical practice 7 reported a 55% prevalence of dysfunction in new women patients attending a general medical clinic. This was not strati ed by underlying disease and is in contrast to a range of 10 ± 35% reported in the general population. 9 Recent research 10 has made progress in evaluating the continuum of female sexuality within a clinical setting. Conceptually, the challenge remains to de ne sexuality without rst imposing a pathological diagnosis. Only then can diminished

2 32 quality of sexual function be appreciated as a side effect of chronic disease or medications. From January 1991 to June 1995, a case ± control study of premenopausal women participants in a rural upstate New York health care system was conducted. The population is almost entirely Caucasian, relatively stable geographically, and well-studied on other health issues. 11,12 An excellent medical record system facilitates patient identi cation and tracking. In this paper we describe the methodology used to evaluate female sexual functioning, hypertension and medication effects in premenopausal women by using ambulatory medical records and community interviews, and the development of a short questionnaire on female sexuality that may be useful in further evaluating this issue. Results are published in a separate paper. 13 Methods Sample selection An a priori sample size of 270 individuals was determined based on the assumption that men and women experience a similar prevalence of sexual dysfunction. 14 Impotence in normal males ranges from 5 to 10% and from 15 to 25% in hypertensive males, both treated and untreated. 5,16 From limited available data, 23% of women who are sexually active and on treatment for mild hypertension experience some sexual dysfunction. 16 Assuming alpha equal to 0.05, power equal to 0.8, and a onetailed rejection region, the sample size required to detect 10% sexual dysfunction in normal individuals vs 25% sexual dysfunction in those with mild hypertension on treatment was 90 subjects per group, using the test of two independent binomial proportions. Three groups were selected to be studied: healthy controls, women with mild unmedicated hypertension, and women with mild hypertension on pharmacological therapy. 17 Due to small numbers, these groups were not matched. Eligible women were identi ed through medical record review. Eligibility criteria were as follows: Caucasian women between 21 and 50 y of age in heterosexual relationships; premenopausal with no history of hysterectomy; no history of pregnancy-induced hypertension or pre-eclampsia; no signi cant medical or recent surgical history; no signi cant psychiatric history, exclusive of situational anxiety or mild depression. Heterosexuality was initially de ned by statement of marital status listed in the medical database in front of the medical record. The medical record was reviewed for any notation indicating change in marital status. Participants were asked in both the self-administered questionnaire (SAQ) and interview about their current marital=partner status. All women indicating involvement in a heterosexual relationship by questionnaire were invited to participate. Women in marriages were not analyzed separately from those not married, but in partnerships due to the focus of the study and small numbers. Women who indicated lack of sexual activity were not included in the analysis. There is no information available on women not sexually active at time of interview as they were excluded from analysis. Healthy controls were selected from those with a diagnosis related group (DRG) for obstetrical delivery, did not carry an International Classi cation of Diseases, 9th edition (ICD 9) code for hypertension and had neither three consecutive blood pressure measurements 140=90 nor an established diagnosis of hypertension recorded in their medical record. Hypertensive cases had an established diagnosis of mild hypertension either by diagnosis in the medical record or three consecutive elevated blood pressure measurements (in the case of discrepancy between ICD 9 code and the medical record abstract, the decision was based on information recorded in the medical record). Recorded medicated or unmedicated blood pressure was < 160=110, but 140=90. Medical records were carefully reviewed for prescribed medications. Those women with diagnosis of hypertension or evaluated blood pressures and recorded medications routinely prescribed for hypertension were classi ed as `medicated hypertension.' Those with no recorded medications were classi ed as `unmedicated hypertension.' After abstracting, the age range for all participants was narrowed to those born between 1950 and 1974 to lessen potential age variability between the healthy controls and hypertensive cases. Only Caucasian women were studied because of the racial homogeneity of the area, and women in homosexual relationships were excluded because of small numbers. Following the records review, a letter outlining the nature of the study and a copy of the study protocol were sent to primary care providers of women who met the inclusion criteria, asking whether there were any contraindications to inviting those women to participate in the study; if an individual's participation was felt to be inappropriate the woman was excluded. A letter was sent to all eligible women introducing the study as concerning women's health and inviting participation, without revealing the study's focus on hypertension. The letter requested a form to be returned in a prestamped envelope, stating whether the woman wished to participate or not, and, in the latter case, why she did not.

3 Data collection and entry The study design included a mailed self-administered questionnaire (SAQ) on sexuality, and a telephone interview to elicit information on past medical=surgical history, medications taken, and sociodemographic factors. Because of the large catchment area and long months of inclement winter weather, in person interviews were deemed impractical and a potential deterrent to participation. Follow-up on initial non-response consisted of telephone re-contact (10 attempts) and r ing of the invitation letter or SAQ. These mailings occurred at three times during the year (spring, summer, and fall) each time that enough charts had been reviewed to justify a research assistant's time. Once the SAQ was returned, the telephone interview was conducted. Subjects not reached via phone were sent follow-up letters. If there was no response after repeated attempts to re-contact, individuals were excluded from the study. Individuals who completed the study (both questionnaire and interview) were sent money orders for $10 as honoraria for their participation. In addition to obtaining supplementary information, the interview served to con rm information obtained from the medical records abstract. When discrepancies occurred, the medical record was rereviewed by one of us (LD). The few individuals with discrepancies that could not be justi ed were deleted from the study. Marital status, race, religion, date of birth, and type of health insurance for all those eligible to participate were obtained from the hospital's nance department data base. All unique identi ers were removed. Data were analyzed by Statistical Applications for the Sciences (SAS). Questionnaire and composite variable development The design of a 64-item SAQ with Likert-type response sets was based on the ndings of previous research on sexuality or on issues in the literature relevant to understanding female sexual functioning, including socioeconomic status and social habits, 18 ± 20 quality of relationship; 21 menstrual, pregnancy, menopausal history; 19 and sexual functioning. 18,20 ± 22 Likert-type responses ranged from ve to seven points; eg never to always. Some of the questions were taken from previously published instruments. The questionnaire was pretested on six women in the community with varying backgrounds to determine if the questions were dif cult to understand or offensive. The interviewer questionnaire inquired about history of hypertension, antihypertensive medication, compliance 19,23 and side effects, 20,24 medical and surgical history, other current medications being taken, and self-reported blood pressure, height and weight. Forty-seven questions on emotional and physiological aspects of sexuality were included in the SAQ; these were classi ed into seven composite variables to reduce the number of outcome measures, limit the number of statistical tests performed, and simplify analyses. Composite variables were formed independently by three investigators (LD, CL and a third female [physician]) using `best clinical judgement.' Questions were then regrouped into composite variables based on DSM IV classi cation 25 and a review of the literature. 10,26 ± 29 Most questions were easily classi ed into one of seven composite variables and had high inter-rater concordance. The few discrepancies were resolved by discussion. To account for the fact that some individuals might not have answered every question, an individual's score for a speci c composite variable was given as a mean response for the number of questions answered, rather than the sum of the responses. The project was approved by the hospital's Research Review Committee and the Institutional Review Board (IRB). In order to handle potential problems effectively, the hospital's public relations of ce and the patient ombudsman were noti ed before potential participants were contacted. A staff psychologist agreed to accept referrals if the investigators deemed this necessary or potentially helpful to a subject. Results Figure 1 summarizes the sampling process. The medical record review identi ed 3312 women; 640 (19%) satis ed inclusion criteria and were cleared for participation by their primary care providers. Of those ineligible, 23 (0.76%) met inclusion criteria but permission to contact was denied, while the remaining 2649 (80%) did not meet at least one inclusion criterion. Of the 640, the response rate (either accept or refuse) to the invitation to participate was 474 (74%). Thirty- ve percent (224) agreed to participate and completed both a SAQ and telephone interview; 110 (49.1%) were healthy controls, 38 (17%) had unmedicated hypertension, and 76 (33.9%) had medicated hypertension. Among those excluded from the study (416), 17 (4%) did not complete the survey, 121 (29%) did not reply, 45 (11%) had moved with no forwarding address, and 233 (56%) refused participation for the following reasons: time constraints (29%), personal issues (28%), not interested (28%), reluctance to participate in research (12%), medical problems (3%). Demographic differences between those who participated in the study and those who either refused to participate or did not respond at all were 33

4 34 Figure 1 Diagram of sampling process. analyzed by w 2 test to determine if the sample was representative (Table 1). None of the factors analyzed reached signi cance. Subjects who could not Table 1 Demographic comparison of enrolled study participants (n ˆ 224) with individuals eligible but not enrolled (nonparticipants) (n ˆ 416) Participants Nonparticipants Marital status* Total Married 77.% 69.% Separated 3.% 3.% Single (includes divorced) 20.% 27.% Insurance* Total Group Program 32.% 35.% Managed care 53.% 42.% Medicare 0.% 1.% Medicaid 7.% 6.% Self 6.% 13.% Other 3.% 4.% Religion* Total Catholic 35.% 39.% Protestant 37.% 38.% Jewish 8.% 7.% Other 5.% 3.% None 15.% 13.% Mean age Total tailed t-test ˆ 1.26, P ˆ 0.21 Mean blood pressure Total =82 130=83 2-tailed t-test ˆ 1.28, P ˆ 0.20 (systolic) 2-tailed t-test ˆ 1.20, P ˆ 0.23 (diastolic) *Data too sparse in certain categories for w 2 test. be located (mail returned with no forwarding address and=or non-working phone) were not included. Age and average blood pressure did not differ signi cantly between groups by t-test (P ˆ 0.21 for age; P ˆ 0.20 for systolic blood pressure and 0.23 for diastolic blood pressure). Table 2 reports the difference in response between the 224 healthy controls and medicated and non-medicated hypertensives deemed eligible by medical record abstracting. Age and blood pressure differed signi cantly (by ANOVA). Demographic variables, by inspection, demonstrate no differences of note except of religious choice (between Catholic and Protestant). A more detailed statistical analysis comparing sociodemographic and sexual function between healthy controls and women with hypertension appears elsewhere. 13 w 2 analyses were used to compare cases and controls on sociodemographic and health variables and will be discussed in a subsequent paper. Each of the 47 questions on the emotional and physiological aspects of sexuality was initially analyzed individually. The seven composite variables derived from these items were desire=aversion, arousal=lubrication, attainment of orgasm, physical pain (dyspareunia only), quality of relationship, satisfaction within the relationship, and frequency of sexual activity. Table 3 reports the distribution of code= recoded sexual response questions among the study group by composite variable, w 2 and P-values. Initial w 2 results were calculated using the full response scale. In order to increase the individual cell size and render more accurate statistics, the response scale was collapsed to `never', `sometimes', `always'.

5 Table 2 Demographic comparison of study participants by healthy controls (n ˆ 110), unmedicated hypertension (n ˆ 38), and medicated hypertension (n ˆ 76) Healthy Unmedicated Medicated controls hypertension hypertension Marital status* Total Married 77.% 74.% 79.% Separated 5.% 3.% 1.% Single (includes divorced) 18.% 24.% 19.% Insurance* Total Group program 29.% 29.% 38.% Managed care 54.% 52.% 52.% Medicare 0.% 0.% 0.% Medicaid 8.% 10.% 3.% Self 6.% 6.% 6.% Other 3.% 3.% 2.% Religion* Total Catholic 45.% 17.% 23.% Protestant 29.% 56.% 43.% Jewish 7.% 6.% 10.% Other 6.% 0.% 7.% None 13.% 22.% 17.% Mean age Total ANOVA F ˆ 25.26, P ˆ Mean blood pressure Total ANOVA F ˆ 64.76, P ˆ (systolic) ANOVA F ˆ 82.33, P ˆ (diastolic) 120=74 132=88 139=89 *Data too sparse in certain categories for w 2 test. w 2 results reported are not age-adjusted. Subsequent analysis deleted an additional 13 women who had undergone surgical or natural menopause. Seven of 47 questions in two composite variables demonstrated signi cant differences between the study groups. In the `Arousal=Lubrication' section, question 31 approached signi cance with hypertensive women (both medicated and unmedicated) having less adequate lubrication than healthy controls. Four additional questions indicated signi cant differences. In question 15, hypertensive women on medication had increased vaginal dryness, while in question 16, women with hypertension had increased vaginal dryness over the last year. Women with hypertension were aroused less by fantasies (question 44) and had more dif culty staying excited (question 51). Three of ve questions in the orgasm composite demonstrated signi cant differences. Women with hypertension reached orgasm less often than usual (question 49), had more dif culty reaching orgasm (question 53), and did not reach orgasm as easily as before (question 54). Most of the composite variables were signi cantly correlated (Spearman's correlation), the principal exceptions being for arousal and quality, arousal and satisfaction, and pain and frequency of sexual activity (Table 4). Stem and leaf plots demonstrated that, with the exception of pain (which showed some right skew), composite variables were grossly normally distributed. (It should be noted that none of the analyses performed relied on the normality of these distributions.) Best questions were generated by selecting those items having the highest Spearman's correlation with their respective composite variables (Table 5). Seven questions thus selected were then used to form an abbreviated questionnaire which may be useful in medical interviews and in research, if replicated in further studies. Discussion The subject of sexual dysfunction as a consequence of drug treatment of hypertension is quite important, because of its obvious impact upon quality of life and as a major factor in noncompliance by male patients. The quality of available data on this subject on men is variable. There are virtually no data on women. Most practitioners virtually ignore this subject when prescribing antihypertensive medication for women, perhaps assuming that the impact upon female sexual function is less than that in males. 30 A strength of this study lies with the fact that it initially examined healthy (except for cases with mild hypertension), sexually active, premenopausal women in an ambulatory outpatient setting. A portion of the survey was community-based, occurring in their homes at the time of the SAQ and telephone interview. The initial 74% response rate 31 may have been due to diligent follow-up by a female interviewer (CS), to female principal investigators (LD, CL), and to the fact that, anecdotally, the study had strong support among the women in the community. The completion rate for the survey and interview was 35%. Based on prior eld research of this nature, 32,33 this completion rate was considered respectable, particularly given the sensitive subject matter and the small, rural communities studied. Seasonal variability in sexual activity and menstrual variability in sexuality 34 were presumed to be randomly distributed due to the cross sectional design of the survey, and the questionnaire and interviewing process taking place in three separate contacts over a year. Non-response bias and self-selection are potential dangers in any study involving behavior especially with the emotional weight of sexuality. Non-participants required more reinforcement (follow-up) to elicit replies to the letter of invitation than did those who chose to participate, which raises the issue of potential reply bias among participants. Careful attempts were made to ensure con dentiality of 35

6 36 Table 3 Coded=recoded sexuality questions classi ed by composite variables I. Quality of relationship w 2 d.f. P-value 7. In general, how would you rate your marriage=partnership? Very happy 5 Happy 4 Neither happy nor unhappy 3 Unhappy 2 Very unhappy 1 8. How does your relationship compare with most of the marriages=partnerships you know? Better than most 3 About the same as most 2 Worse than most 1 9. How would you rate your marriage or partnership at present? No problems 7 Very minor problems 6 Minor but persistent problems 5 Major problems, but divorce or separation not being considered 4 Major problems and divorce or separation being considered 3 Separation has occurred 2 Divorce in progress 1 II. Desire=aversion Most times Half the time Sometimes Never Always In the last year, in your sexual partnership, how often: 25. Do you relax during sex? Are you interested in sex? Do you regard yourself as having any problems with sex? Are you turned off= nd sex distasteful? In the last year, do you nd that you: 35. Enjoy sex less than you used to? Have less interest in sex than you used to? Think about sex less than you used to? Are attracted to any person(s) other than your partner? Reject the advances of your partner? In the last year, using the methods of stimulation described in questions 40 through 45: 47. Did you choose these method(s) because of a desire for self-stimulation=masturbation? III. Arousal=lubrication Regarding your health: 15. Do you have vaginal dryness? c No Ð Go to question 17 2 Yes Within the last year, have you noticed increased c vaginal dryness? No 2 Yes 1 In the last year in your sexual partnership, how often: 31. Do you have enough natural lubrication (wetness) during sex? Always Most Half times the time Sometimes Never a Most times Half Never the time Sometimes Never Tried Always In the last year, when you have been alone, have you been aroused by: 40. Books Films Magazines=pictures Dreams Fantasies b 45. Self-stimulation=masturbation In the last year, have you: 50. Had more dif culty getting excited? Had more dif culty staying excited? b Table 3 continued

7 Table 3 continued IV. Orgasm Most times Half the time Always Never In the last year, using the methods of stimulation described in questions 40 through 45: 46. Did you reach orgasm=climax=come? In the last year, have you: 49. Reached orgasm=climax=come less often than usual? b 52. Reached orgasm=climax=come too quickly? (quicker than before) 53. Had more dif culty reaching orgasm=climaxing= b coming? 54. Found you cannot reach orgasm=climax=come as b easily as before? V. Physical pain In the last year, in your sexual partnership, how often: 32. Is intercourse physically painful? In the last year, have you: 55. Found intercourse physically more painful? VI. Activity=Frequency 23. On average, how often do you and your partner engage in any type of close sexual contact? More than three times a week 6 1 ± 3 times a week 5 More than three times a month 4 1 ± 3 times a month 3 More than three times a year 2 1 ± 3 times a year Has the frequency of this contact changed during the last year? No change 3 Yes, increased greatly 5 Yes, increased slightly 4 Yes, decreased slightly 2 Yes, decreased greatly 1 Always Most times Half the time Sometimes Never 34. Are having sex less often than you used to? On average, how many times per week=month do you engage in self-stimulation (masturbation)? More than three times a week 7 1 ± 3 times a week 6 More than three times a month 5 1 ± 3 times a month 4 More than three times a year 3 1 ± 3 times a year 2 Never Has this changed during the last six months? No change 3 Yes, increased greatly 5 Yes, increased slightly 4 Yes, decreased slightly 2 Yes, decreased greatly 1 VII. Satisfaction Most Half Always times the time Sometimes Never In the last year, in your sexual partnership, how often: 29. Do you have too little foreplay before sex? Do you have too little gentleness during=after sex? Is intercourse emotionally painful? In the last year, using the methods of stimulation described in Questions 40 through 45: 48. Did you nd the same degree of enjoyment using these method(s) as when you had sex with a partner? In the last year, have you: 56. Found intercourse emotionally more painful? In the last year, in your sexual partnership, how often: 57. Does your partner have different sexual habits=preferences than you? Some times Does your partner choose an inconvenient time for sex? Table 3 continued 37

8 38 Table 3 continued 59. Does your partner initiate caressing=petting=sex? Do you and your partner caress=pet=kiss during the day without having sex=intercourse? 61. Do you and your partner caress=pet=kiss during the evening= night without having sex=intercourse? 62. Has your partner had dif culties with sex in the past six months (eg impotence, premature ejaculation, low sex drive)? a Approaches signi cance; b P < 0.05; c P < Table 4 Spearman correlation coef cient matrix for seven coded recoded composite variables Quality Desire Arouse Orgasm Pain Freq Satisfy Quality Desire Arouse Orgasm Pain Freq Satisfy Table 5 Best questions generated by highest correlation to composite Spearman's correlation coef cient 1. In the last year, in your sexual partnership, how often is intercourse physically painful? In general, how would you rate your marriage=partnership? In the last year, have you had more dif culty reaching orgasm=climaxing=coming? In the last year, do you nd that you have less interest in sex than you used to? In the last year, do you nd that you are having sex less often than you used to? In the last year, when you have been alone, have you been aroused by fantasies? In the last year, in your sexual partnership, how often do you have too little gentleness during=after sex? 0.65 participants and smooth handling of concerns or complaints from subjects, of which there were very few. There were several limitations. The study population was limited to Caucasian woman because of the underlying demographics of rural upstate New York. The high prevalence of hypertension among African ± Americans and other ethnic groups calls for research of this kind in other settings. Participants were not followed over time and were not asked to keep daily sexual diaries, used in another study, 8 which might have improved the accuracy of reporting, but would have introduced bias against spontaneous answers. Their sexual partners were not interviewed, which might have provided a different dimension to participants' responses. Questions on sexuality in the SAQ were not validated externally by other questionnaires reported in the literature or internally by test ± retest on our population, although they were derived from questionnaires in the published literature. 18 ± 24 The issue of evaluating the questions for validity is dif cult as there are no set criteria or `gold standards' 35 for measuring sexuality, a quality of life issue which is conceptually subjective and dependent upon the perception of the individual. Each question=composite variable was examined for face and content validity by the authors. 31 The only alternative to strengthening face validity would have been to increase the panel of experts. A test ± retest examination of reliability was beyond the scope of this study. Further research into this area should help to further de ne both validity and reliability. Additionally, participants answered questions on a one to ve or seven scale; however, some questions were negatively worded which may have in uenced responses. Medications and dose information were self-reported; con rmation by pharmacy records would have improved this aspect but was not feasible. Although current blood pressure readings could not be obtained, averages of the last three blood pressures recorded on separate visits in each medical record were used. Due to small study numbers, groups were not matched with regard to the number and class of antihypertensive agents taken. All patients meeting study criteria were included regardless of dose of antihypertensive agent taken, whether or not blood pressure was controlled on medication, and whether or not patients were compliant with medication. However, all attempts were made to include individuals only on monotherapy. Concomitant use of alcohol, other medications potentially impacting on sexual response, and oral contraceptives

9 were not considered in the analysis, although these data were collected either during medical record abstraction or in the SAQ=telephone interview survey. To have included these data in the analysis would have depleted the numbers in each cell, thus rendering the results uninterpretable. Clearly, future studies should consider these factors should their sample sizes be suf cient to support appropriate analysis. Initial unadjusted w 2 analysis suggested that women with hypertension, both medicated and unmedicated, many have increased dif culty with adequate lubrication during intercourse and in achieving orgasm; this is discussed further in another paper. 13 The brief questionnaire developed during this study may provide a means to assess quickly the physiological and emotional components of sexual functioning with respect to hypertension, within a population or clinical setting. Further work in this area will determine the brief questionnaire's validity, its ability to detect problems with sexual functioning in women with chronic disease, such as hypertension, diabetes, and depression, and generalizability to other population subgroups. This study establishes a method to survey healthy (except for mild hypertension), sexually active, premenopausal women in an ambulatory outpatient setting to address the effects of hypertension and pharmacotherapy on sexuality. It attempted to provide a continuum of sexual normalcy rather than conform to a de nition of sexual dysfunction. It effectively employed self-administered questionnaires and telephone interviews to elicit candid answers to sensitive questions. With recent advances in research on female sexuality, population-based studies on the effects of chronic disease, medications and sexuality are increasingly feasible. By improving and expanding existing methodologies and questionnaires, population-based surveys and clinical trials may make major contributions to understanding the effect of chronic disease and therapy on female sexuality. Acknowledgements This project was funded by an E. Donnall Thomas Research Award T and by the Research Institute, Bassett Healthcare, One Atwell Road, Cooperstown NY The authors wish to thank the E. Donnall Thomas Resident Research Committee for its support through the protocol development and data collection phase of this project. Ms Mary Muller and Mr Chris Cawley are thanked for their considerable effort in typing and manuscript production, and Ms Catherine Mason for expert editorial assistance. Ms Sara Clark and the medical records staff gave invaluable assistance and support as did Ms Cindy St John of Information Services. Dr Duncan thanks Dr DN Bateman. Some of the results from this paper were presented at the American Heart Association annual meeting, Orlando, Florida, November 1997; The Fifth International Congress of Behavioral Medicine, Copenhagen, Denmark, August 1998; and in abstract form in New Perspectives in the Management of Female Sexual Dysfunction, American Urologic Association, Boston, Massachusetts, October References 1 Duncan L, Bateman DN. Sexual function in women: do antihypertensive drugs have an impact? Drug Saf 1993; 8: 225 ± Prisant LM et al. Sexual dysfunction with antihypertensive drugs. Arch Int Med 1994; 154: 730 ± Bateman DN. Drugs and sexual function. Adverse Drug React Bull 1980; 80: 308 ± Lewis C, Duncan LE, Ballance D, Pearson TA. Sexual dysfunction in hypertensive women: uncommon or understudied? Am J Hypertens 1998; 11: 6; part 1, 733 ± Bulpitt CJ, Dollery CT, Carne S. Changes in symptoms of hypertensive patients after referral to hospital clinic. Br Heart J 1976; 38: 121 ± Grimm Jr RH et al. Long-term effects of sexual function of ve antihypertensive drugs and nutritional hygienic treatment in hypertensive men and women, treatment of mild hypertension study (TOMHS). Hypertension 1997; 29: 1; part I, 8 ± Ende J, Rockwell S, Glasgow M. The sexual history in general medicine. Arch Int Med 1984; 144: 558 ± Plouffe L. Screening for sexual problems through a simple questionnaire. Am J Obstet Gynecol 1985; 151: 166 ± Laumann EO, Gagnon JH, Michael RT, Michaels S. The Social Organization of Sexuality; Sexual Practices in the United States. Chicago: The University of Chicago Press: Chicago, Rosen RC, Taylor JF, Leiblum SR, Bachman GA. Prevalence of sexual dysfunction in women: results of a survey study of 329 women in an outpatient gynecological clinic. J Sex Marital Ther 1993; 19: 171 ± Bowlin SJ et al. Validity of cardiovascular disease risk factors assessed by telephone survey: the Behavioral Risk Factor Survey. J Clin Epidemiol 1993; 46: 561 ± Bowlin SJ et al. Reliability and changes in validity of selfreported cardiovascular risk factors using dual response: the Behavioral Risk Factor Survey. J Clin Epidemiol 1996; 49: 511 ± Duncan LE, Lewis C, Jenkins P, Pearson TA. Does hypertension and its pharmacotherapy affect the quality of sexual function in women? Am J Hypertension 2000; 13: 640 ± Moss HB, Procci WR. Sexual dysfunction associated with oral antihypertensive medication: a critical review of the literature. Gen Hosp Psychiatry 1982; 4: 121 ± Smith PJ, Talbert RL. Sexual dysfunction with antihypertensive and antipsychotic agents. Clin Pharm 1986; 5: 373 ± Poloniecki J, Hamilton M. Subjective costs of antihypertensive treatment. Hum Toxicol 1985; 4: 287 ± The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Int Med 1997; 157: 2413 ± Kelsey JL, Thompson WD, Evans AS. Measurement 1. Questionnaires in Methods in Observational Epidemiology. Oxford University Press: New York, 1986, pp 309 ± Osborn M, Hawton K, Gath D. Sexual dysfunction among middle aged women in the community. Br Med J 1988; 296: 959 ±

10 40 20 Bulpitt CJ, Dollery CT. Side effects of hypotensive agents evaluated by a self-administered questionnaire. Br Med J 1973; 3: 485 ± Croog SH et al. The effects of antihypertensive therapy on the quality of life. New Engl J Med 1986; 314: 1657 ± Frank E, Anderson C, Rubinstein D. Frequency of sexual dysfunction in `normal' couples. New Engl J Med 1978; 299: 111 ± Leeton J, McMaster R, Worsley A. The effects on sexual response and mood after sterilization of women taking longterm oral contraception: results of a double-blind cross-over study. Aust NZ J Obstet Gynaecol 1978; 18: 194 ± Bulpitt CJ, Dollery CT, Carne S. A symptom questionnaire for hypertensive patients. J Chronic Dis 1974; 27: 309 ± American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders 4th edn. American Psychiatric Association: Washington, DC, Taylor JF, Rosen RC, Leiblum SR. Self-report assessment of female sexual function: psychometric evaluation of the brief index of sexual functioning for women. Arch Sex Behav 1994; 23: 627 ± Rosen RC, Leiblum SR. Treatment of sexual disorders in the 1990s: an integrated approach. J Consult Clin Psychol 1995; 63: 877 ± Masters WH, Johnson VE. Human Sexual Response. Little Brown: Boston, Kaplan HS. Hypoactive sexual desire. J Sex Marit Ther 1977; 3: 3±9. 30 Bauer GE, Baker J, Hunyan SN, Marshall P. Side-effects of antihypertensive treatment: a placebo-controlled study. Clin Sci Mol Med 1978; 55: 341S ± 344S. 31 Dillman DA. Mail and Telephone Surveys: The Total Design Method. John Wiley: New York, Bickman L, Rog DJ (eds). Handbook of Applied Social Research Methods. Sage publications: Thousand Oaks, CA, Aday LA. Designing and Conducting Health Surveys. Jossey- Bass: San Francisco, Hodge RH et al. Sexual function of women taking antihypertensive agents: a comparative study. J Gen Int Med 1991; 6: 290 ± Wassertheil-Smoller S, Davis BR. Quality of life and antihypertensive therapy. Prim Cardiol 1992; 18: 54 ± 60.