- Email: [email protected]
Successful recruitment strategies for women in postpartum mental health trials
Journal of Psychiatric Research 37 (2003) 117–125 www.elsevier.com/locate/jpsychires
Successful recruitment strategies for women in postpartum mental health trials Kathleen S. Peindla,*, Katherine L. Wisnerb a
Department of Psychiatry & Human Behavior, Thomas Jefferson University, 833 Chestnut Street, Suite 210-E, Philadelphia, PA 19107, USA b Western Psychiatric Institute & Clinic, University of Pittsburgh, Pittsburgh, PA 15213, USA Received 3 May 2002; received in revised form 7 September 2002; accepted 29 October 2002
Abstract Recruiting women into research protocols allows investigators to examine the efficacies of treatments and to study other outcomes among women with mental illnesses. However, achieving recruitment goals has been difficult for researchers. The objective of this study was to examine the success rates of different strategies for recruiting women into clinical trials for the prevention and treatment of postpartum major depression. This is a descriptive study designed to examine which recruitment efforts, over a 4-year period, yielded women who participated in the studies. It was conducted in an outpatient clinic affiliated with a major urban teaching hospital where women of childbearing age sought treatment for affective disorders. All women (n=589) who called about our research studies were systematically screened. The women were either pregnant or in the postpartum period. Our results show that in both studies, obstetricians and other healthcare professionals referred a large number of women. These referrals yielded relatively high rates of participation. However, other methods were differentially successful for the two studies: media appearances and advertising resulted in almost 50% of the screening calls for the prevention study, whereas mass mailings and other promotional materials were more effective for the treatment study. Those materials contributed 44% of screening calls for the treatment study. Our conclusions were that women were most likely to enter the research studies when referred by an obstetrician or other professional. # 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction Before 1977, women of childbearing age were excluded for Phase I & Phase II clinical trials. They could only be included if animal studies determined there were no effects of the outcome on the fetus or female fertility. The paradox is that all women were excluded if they were of childbearing age. Only AIDS and cancer studies included women (Merkatz et al., 1993). In the 1980s, public health leaders drew attention to the inequitable representation of women and minorities in National Institutes of Health-funded studies. The GAO reviewed participation of women and recommended that the FDA needed to ensure gender representation in prescription drug testing. Their study also found that reports from trials that included women did not examine gender differences in outcomes [GAO Pub# GAO/ * Corresponding author. Tel.: +1-215-503-2817. E-mail address: [email protected] (K.S. Peindl).
HRD-93-17, 1993). New guidelines from the FDA in 1993 stated that women should be included in ‘‘reasonable’’ numbers in all clinical trials. The National Institutes of Health (NIH) published An Outreach Notebook (La Rosa et al., 1997) that offers guidelines for the inclusion of women and minorities in research protocols. To date, NIH now mandates these guidelines. A recent article published the results about the representation of women in randomized trials of acute coronary syndromes (Lee et al., 2001). The authors reported that enrollment of women into trials has increased by 25% but this is still below the proportion of women with myocardial infarctions in the general population, which is 43%. As of 2001, women were still under represented in breast cancer trials (Evans et al., 2001). The lifetime risk of breast cancer in the United Kingdom is 40%, but the enrollment into three separate trials for breast cancer has been 10%. Women who were eligible did not enter the trials and 60% chose to be screened. Even with women being underrepresented in
0022-3956/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved. PII: S0022-3956(02)00086-9
118
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
research studies, great strides in understanding women’s health issues have occurred over the past decade. Two large trials, The Women’s Health Initiative (WHI) and The Postmenopausal Estrogen/Progestin Interventions Trial (PEPI) are studies specifically designed to address inequities in women’s health relating to heart disease, cancer, osteoporosis and the effects of hormone replacement on heart disease risk in postmenopausal women. Both trials have had success in recruiting women. Recruitment efforts for the WHI included mass mailings, community outreach, distribution of brochures at many community events, and professional referrals. The enrollment of women into each of these trials was highly successful (Lewis et al., 1998; Johnson et al., 1995). Several investigators have published reports on strategies for recruiting women. Lovato et al. (1997) published a summary of recommendations for the recruitment of women into clinical trials, which included flexible appointment schedules, provisions for daycare, placement of advertisements that target women, and recruitment for the trial within female-dominated practices. Five well-planned strategies were used to recruit women into a 3-year study of caregivers for chronically disabled adults (Patrick et al., 1998). In this trial, the number of participants required to detect differences among the study groups was 800. Implementation of the five strategies allowed for successful recruitment of 844 women. First, the research staff contacted 48 formal service agencies that served women who might be eligible for the study. This strategy accounted for 29.3% of the total sample; 85.4% of those initially recruited completed the study. Second, the research staff presented material about the study to support groups. This strategy accounted for 9.2% of the participants, with a 97.5% retention rate throughout the study. Third, enrolled study participants recommended other women to the researchers, who contacted those potential participants. This strategy accounted for 37.3% of the total sample; their study completion rate was 88.9%. Fourth, the investigators used media and promotional materials (press releases, radio and newspaper advertisements, and distribution of brochures and flyers) to inform potential participants. This method yielded an unknown percentage of the total sample. The fifth and final strategy used a modified demographic sampling unit, which resulted in 0.7% participation and a moderate completion rate of 79.3%. Several investigators have had relative success in recruiting women through mass mailings or face-to-face screening (Lewis et al., 1998; Gilliss et al., 2001; Baines, 1984). In one study, investigators needed subjects for a three-site trial to reduce dietary fat intake in postmenopausal women. By mass mail, they contacted 210,000 women, 11.5% were randomized into the trial (Lewis et al., 1998). The second study was a 5-year investigation of the health of women mid-life through menopause transition (Gilliss et al., 2001). In this study,
face-to-face screening yielded the highest percentage of participation (74%). In the third study, investigators in a Canadian breast cancer screening program had success by following up mass mailings with telephone contacts (Baines, 1984). Of the women contacted by this method, 26.4% entered the study. However, the number of participating women (from mass mailing and phone contact) was less than 1% of the total sample needed for screening. In all three studies, other recruitment methods yielded small numbers of enrolled women. Another group of investigators utilized the strategy of recruiting women from a primary care network into a genetic counseling study for breast cancer risk (Hlemes et al., 2000). Mass mailing used for the recruitment lasted 11 months. Of the women contacted, 8.6% entered the study. The investigators reported that even fewer women wanted to remain in the protocol after learning of future commitments of the research study. Some investigators have examined women’s reasons for agreeing to participate in a research study. Krieger et al. (2001) held focus group sessions for cancer patients who had been asked to participate in a research study. They asked attendees what factors affected their decision to participate. The element with the most positive impact for participants was a pre-enrollment discussion with research staff about the time commitment involved in the study. This paper describes our efforts to recruit women into two clinical trials. The first trial was for the prevention of postpartum major depression and the second trial was for the treatment of postpartum depression. In our studies, we encountered two additional recruitment issues because of our focus on the mental health of pregnant and postpartum women. First, it was necessary to rely on other medical specialists’ practices to refer women to our studies. Other investigators have also described the problem of referrals from other clinical departments (Hlemes et al., 2000). Second, we faced the obstacle of a lack of awareness about mental illness among pregnant and postpartum women. Our two studies were approved by the IRB of the University Hospitals of Cleveland. The first trial was a randomized, double-blind, clinical trial that determined the efficacy of nortriptyline compared to placebo for the prevention of postpartum depression. In this trial, women were healthy during the current pregnancy but had previously experienced an episode of postpartum depression (Wisner et al., 2001). The second trial is a randomized, double-blind, controlled clinical trial comparing nortriptyline versus sertraline for the treatment of postpartum depression.
2. Method We screened all calls to our women’s mental health clinic, which specializes in the treatment of affective
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
disorders among women of childbearing age. We recorded information on a standardized screening form (see Appendix). The form included questions about pregnancy or postpartum status, demographic information, insurance coverage, and original source of information about the study. When the caller was referred by a healthcare professional, we recorded the professional’s name, phone number, and specialty (e.g. obstetrician, mental health therapist, midwife). Other information sources included promotional materials, media appearances, and advertisements. 2.1. Treatment study A mailing was sent by first-class mail to all women who delivered in the previous month at the University Hospitals of Cleveland. On average, 450 infants were delivered each month. The mailing included a letter of congratulations on the birth of a baby, information about postpartum depression and our clinic, our expression of concern for their postpartum well-being, and a flyer about our postpartum treatment study. Both the letter and flyer were approved by the IRB. Two pamphlets were also sent: Mom never had a chance to be involved in medical research, published by the Society for Women’s Health Research (2000), and Volunteering for a clinical trial, published by CenterWatch (1999). We enclosed a self-rating scale for postpartum depression; the Edinburgh Postnatal Depression Scale (EPDS, Cox et al., 1987), with instructions to call us if they scored 12 or higher, felt depressed, or had questions. We encountered no difficulty in women completing the screen. They were willing to complete it as a check of their emotional wellness. We obtained the names and mailing addresses of the women from the hospital database after IRB and hospital approval. The mailing list was also available to other companies who direct mail to mothers with newborns. The women in the mailing did not have, to our knowledge, any mental illness but some were at risk for having an onset of major depression. We explained that risk in the enclosed letter. Today, it would be unlikely that an IRB would approve this type of mailing because of the breach in confidentiality of these patients. Study brochures were sent to all obstetricians in the Cleveland area, with a request to place them in their offices for patients. The research staff also visited offices and requested permission to place flyers and brochures in the waiting rooms. We found that the research staff placing study information in the offices was most successful. With the obstetricians’ packet of brochures, we included information about our clinical practice and research studies. In an enclosed letter to the doctors, we asked to schedule a meeting to present information about the clinical practice and our studies. We presented to three out of ten large obstetrical practices that were willing to meet with us. We also contacted 150
119
obstetrics and gynecology physicians by mail. Twentyfour (16%) of the contacted physicians referred to our studies. The Principal Investigator (KLW) reached other obstetricians by presenting our research projects at meetings of the local Obstetrics and gynecology Society. Several members of the research staff gave talks to family medicine residents, nurse midwives, psychiatric and obstetric residents, and the nursing department within the university. At each presentation, we provided flyers and brochures about our studies and our clinical services. In addition, the research staff placed brochures and flyers in libraries, hospitals, and clinical practice buildings throughout the city. We also utilized the media for recruitment purposes. The PI (KLW) and several clinicians served as guests on local television talk shows about postpartum depression and other mental illnesses that are more prevalent among women. We placed advertisements about the studies and our clinic in newspapers, on the radio and the scroll bar of a local television weather station. 2.2. Prevention study Our recruitment strategies for the prevention study were similar to those for the treatment study. A mass mailing was sent in a pre-admission packet to all women who delivered at University Hospital. We did not have access to the names and addresses for this mailing that was done by hospital personnel. It included a letter, a brochure about the study, and an information sheet on postpartum depression. These enclosures were approved by the IRB of the hospital. Brochures, posters with informational tear-offs, and flyers were placed in obstetrics offices throughout the city. We presented to the local Obstetrics and gynecology Society, midwives, nurses, and obstetrics and gynecology offices. Mass media presentations included morning television talk shows; a segment on PrimeTime Live (ABC); and cable television, radio, and newspaper advertisements. For the prevention study, we also utilized a face-toface screening process. One of our research staff (nurse) gained access to a large hospital-based obstetrics clinic and screened pregnant women for history of postpartum major depression. The nurse was time-limited and her clinical obligations permitted only 1 day per week to screen. Because of the strict eligibility requirements for the study, we did not find any eligible women. If we would have had a full-time staff member who was able to screen daily, we may have recruited women for the study.
3. Results Table 1 shows the number of contacts generated by each type of information source for each study. Professional
120
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
Table 1 Number of women screened, by information source and study Information source
Number of contacts per study (%)
Table 2 Number of women participating in treatment study, by information source Information source
Number screened Total=283
Number participating (% of total participants) N=77
(48) (28) (7) (4) (3) (4) (2)
Professional referrals—total Obstetrician Nurse Counselor/social worker Pediatrician Psychiatrist Midwife
136 79 19 11 9 11 7
38 21 7 2 5 2 1
(49) (27) (9) (3) (6) (3) (1)
16 5 11
(6) (2) (4)
Other referrals—total Insurance advocate Family member or friend
16 5 11
8 4 4
(10) (5) (5)
(6) (3) (3) (0) (0)
125 103 19 1 2
(44) (36) (7) ( <1) ( <1)
Promotional materials—total Mass mailing (postpartum) Brochures/flyers Hospital video Website
125 103 19 1 2
30 25 5 0 0
39 (32) (6) (0) (0)
64 45 11 8
(21) (15) (4) (3)
2 2 N/A 0
( <1) ( <1)
Local television talk shows Radio advertisements
2 4
0 1
(0) (1)
Media advertisements—total Cable television Radio Newspaper
88 65 2 21
(29) (21) ( <1) (7)
4 0 4 0
(1) (0) (1) (0)
Face-to-face screening Ohio Task Force
44 1
(14) ( <1)
N/A 0
(0)
Prevention N=306
Treatment N=283
Professional referrals —total Obstetrician Nurse Counselor/social worker Pediatrician Psychiatrist Midwife
83 58 5 2 0 12 6
(27) (19) (2) ( <1) (0) (4) (2)
136 79 19 11 9 11 7
Other referrals—total Insurance advocate Family member or friend
8 0 8
(3) (0) (3)
Promotional materials—total Mass mailing Brochures Hospital video Website
18 8 10 0 0
Media appearances/news—total Local television talk shows PrimeTime Live (ABC) Other television news
(0)
referrals resulted in a high number of completed screens for both studies. In the treatment study, the mass mailing was particularly effective in generating contacts, whereas in the prevention study, the cable television advertisements resulted in more contacts. 3.1. Treatment study We screened 283 women for the postpartum depression treatment study over a 4-year period. Thirty-five percent (n=99) of the women refused the study after the initial screen. Among those who refused, a physician referred 28% of the women and 38% received the postpartum letter. Women had an initial evaluation to determine eligibility. Among those women (n=107), 38 percent of the women were ineligible. Seventy-seven women (13%) agreed to participate in the study. One woman refused medication after randomization into the trial. Table 2 shows the number of actual study participants generated by each type of information source for this study. The greatest proportion of women who contacted
us (48%) were referred by healthcare professionals—a group comprised of obstetricians, nurses, counselors, social workers, pediatricians, psychiatrists, and midwives. Women referred by these professionals made up almost half of our study participants. The second largest number of phone contacts to the clinic came from the mass mailing, making up 36% of the women screened and ultimately 32% of our study participants. Interestingly, although the number of screening contacts generated by insurance advocates was small (n=5), a high percentage of these women (80%) entered the study. The insurance companies had no stake in the research. All visits and medications were free to the participant, but two women within the insurance companies were advocates for the study and made referrals to trial. The staff salaries and medications were furnished from grant expenses. 3.2. Prevention study We screened 306 women for the postpartum depression prevention study over a 3.5-year period. Forty percent of the women screened did not fit criteria for entry into the trial and 20% of the women refused participation after the phone screen. Among the screened women, 121 were eligible for the trial and fifty-seven women entered the study. Five women declined medication after randomization and one woman became manic during the first week postpartum. For this study, the participation rate was 42% among eligible women. Referrals from healthcare professionals again yielded high screening and participation rates. However, the
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
results for other sources were slightly different from those for the treatment study. Table 3 shows the number of study participants generated by type of information source for this study. Media appearances and advertisements generated a large volume of screening calls. However, they did not always convert to high rates of study participation. Appearances on local television talk shows resulted in 45 screening calls, but these women ultimately comprised only 7% of the sample. Advertisements on cable television were more successful, resulting in 65 screening calls and 21% of study participants. Other recruitment strategies yielded very small numbers of women entering the study; notably, the face-to-face screening process generated no study participants.
4. Discussion The limitation of this study is that the two trials were specifically designed for treatment and prevention of major depression in postpartum women. We needed to be sensitive to mental health issues among the physicians and the women that we were seeking to enter the trials. The trials were conducted in an academic setting, but we recruited women throughout the metropolitan Table 3 Number of women participating in prevention study, by information source Information source
Number screened Total=306
Number participating (% of total participants) N=57
Professional referrals —total Obstetrician Nurse Counselor/social worker Psychiatrist Midwife Family member/friend referrals
83 58 5 2 12 6 8
25 20 2 0 1 2 2
(44) (35) (4) (0) (1) (4) (4)
Promotional materials—total Mass mailing (pre-admission) Support group mailing Brochures/posters
25 8 7 10
9 5 0 4
(16) (9) (0) (7)
Media appearances/news—total Local television talk shows PrimeTime Live (ABC) Other television news
64 45 11 8
5 4 1 0
(9) (7) (2) (0)
Media advertisements—total Cable television Radio Newspaper
82 65 2 15
16 12 0 4
(28) (21) (0) (7)
Face-to-face screening Ohio Task Force
44 1
0 0
(0) (0)
121
community. Recruitment for both studies depended largely upon referrals from physicians and other professionals outside the academic community. At first, referrals from physicians were difficult. We found that at least one year was necessary to develop relationships within the community and to establish a referral network of obstetricians. When developing these relationships, we emphasized our interest in their patients’ wellbeing rather than placing emphasis on mental illness. It took time to build their trust and we established trust with successful treatment of their patients or having one of their patient (by self-referral] enter a study and successfully complete it. Our report shows that the relationship that the patient has already established with her healthcare provider encourages research participation when a referral is made. For successful recruitment, strategies should be planned and implemented before beginning the study, with follow-up plans formulated to keep the studies fresh in the minds of health professionals and to replenish brochures and flyers on a monthly basis. During the planning stage, the research team should realistically address the labor and time commitments of the proposed methods of recruitment. For example, mass mailings for 450 deliveries per month take 4 h to complete. Monthly flyer development and copying of enclosures in the mailing takes 6 h. Delivering flyers and brochures to obstetric and gynecology offices takes twice a month with deliveries made twice a week. These commitments may also include after-office-hour presentations for television and radio shows. Media outlets are attractive for recruitment purposes because of the wide audience they reach. However, we had mixed results with our media-related strategies. Although they generated a high number of calls, they yielded low rates of participation. Media advertisements can be costly, so it is wise to plan carefully the timing and placement of any ads. For example, we found that our advertisements on the cable television weather channel (for the prevention study) were successful only during the winter months, when people were more concerned with weather changes. We had no response to this recruitment strategy during the summer. Referrals by health care professionals were our best source of potential study candidates (49, 44% participation rates). Although time consuming, mass mailings gave us 32% of the sample in the treatment study. Careful use of TV (in our case, cable weather channel) did give us participants for the prevention study (21%). We cannot measure the impact of the educational programs presented on TV or radio talk shows about the importance of women’s mental health. Any information given to the public helps women come for treatment as in the case of breast cancer awareness. However, more information needs to be disseminated about the importance of research to the knowledge and treatment of
122
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
psychiatric disorders among women. Without their trust in the research team and in the research endeavor, we will still encounter difficulties in recruiting women into clinical trials. We make the following suggestions for recruitment of women into mental health research: 1. Establish relationships with psychiatric and primary care physicians, and obstetricians. Depending on the type of study, different practices can be targeted for education about the research and women’s mental health. In our studies, we contacted obstetrics and gynecology practices, midwives, pediatricians, obstetrical nurses, and social workers about the importance of the studies to the treatment and prevention of postpartum depression. Family Practice physicians and insurance companies are another potential source for referrals. 2. Encourage primary care physicians to be ambassadors for the project by educating them about mental illness and the importance of your research to the well-being of their patients. Provision of information to physicians about their patients’ progress in research protocols is imperative to establish trust. 3. Inform the research staff of the time commitments involved in developing and implementing
recruitment strategies. This time commitment includes developing strategies through regular weekly meetings, calling offices and making appointments to meet with specialists and their staff, arranging to deliver flyers and brochures to specialists’ offices, creating flyers for mass mailings, obtaining information (addresses) for mass mailings, and presenting informative talks to departments, societies, and public health officials about research and women’s mental health issues. These presentations may be after regular office hours. 4. Educate the public—which includes potential research participants—not only about the mental health issue being studied, but also about the overall importance of research in improving mental health care. The information can be disseminated by various methods; for example, we used mass mailings to provide women with information about postpartum depression and participating in clinical trials.
Acknowledgements Funding for the project came from National Institutes of Mental Health MH53735 and MH57102.
Appendix follows on next page.
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
Appendix
123
124
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
References Baines CJ. Impediments to recruitment in the Canadian national breast cancer screening study: response and resolution. Controlled Clin Trials 1984;5:129–40. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry 1987;150:782–6. Evans DGR, Lalloo F, Shenton A, Boggis C, Howell A. Uptake of screening and prevention in women at high risk of breast cancer. Lancet 2001;358:889. Gilliss CL, Lee KA, Gutierrez Y, Taylor D, Beyene Y, Neuhaus J, et al. Recruitment and retention of healthy minority women into community- based longitudinal research. J Women’s Health & Gender-based Medicine 2001;10:77–85. Hlemes AW, Bowen DJ, Bowden R, Bengel J. Predictors of participation in a primary care physician network. Cancer Epidemiology 2000;9:1377–9. Johnson S, Mebane-Sims I, Hogan PE, Stoy DB. Recrutiment of
postmenopausal women in the PEPI trial. Postmenopausal Estrogen/Progestin Interventions. Controlled Clin Trials 1995;16(4):20S– 35S. Krieger N, Ashbury F, Cotterchio M, Macey J. A qualitative study of subject recruitment for familial cancer research. Ann Epidemiology 2001;11:219–24. La Rosa JH, Parron D, Alvarado M. Outreach notebook for the NIH guidelines on inclusion of women and minorities as subjects in clinical research. National Institutes of Health publication No. 97-4160; 1997. Lee PY, Alexander KP, Hammill BG, Pasquali SK, Peterson ED. Representation of elderly persons and women in published randomized trials of acute coronary syndromes. JAMA 2001;286:708–13. Lewis CE, George V, Fouad M, Porter V, Bowen D, Urban N. Recruitment strategies in the Women’s Health Trial: feasibility study in minority populations. Controlled Clin Trials 1998;19:461–76. Lovato LC, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruitment for controlled clinical trials: literature summary and annotated bibliography. Controlled Clin Trials 1997;18:328–57.
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125 Merkatz RB, Temple R, Sobel S, Fieden K, Kessler DA. Women in clinical trials of new drugs—a change in Food and Drug Administration Policy. NEJM 1993;329:292–6. Mom never had the chance to be involved in medical research [pamphlet]. Society for Women’s Health Research; Washington, DC; 2000. Patrick JH, Pruchno RA, Rose MS. Recruiting research participants:
125
a comparison of the costs and effectiveness of five recruitment strategies. The Gerontologist 1998;38:295–302. Volunteering for a clinical trial [pamphlet]. Boston, MA: CenterWatch, 1999. Wisner KL, Perel JM, Peindl KS, Hanusa BH, Findling RL, Rapport D. Prevention of recurrent postpartum depression: a randomized clinical trial. J Clin Psychiatry 2001;62:82–6.
Successful recruitment strategies for women in postpartum mental health trials Kathleen S. Peindla,*, Katherine L. Wisnerb a
Department of Psychiatry & Human Behavior, Thomas Jefferson University, 833 Chestnut Street, Suite 210-E, Philadelphia, PA 19107, USA b Western Psychiatric Institute & Clinic, University of Pittsburgh, Pittsburgh, PA 15213, USA Received 3 May 2002; received in revised form 7 September 2002; accepted 29 October 2002
Abstract Recruiting women into research protocols allows investigators to examine the efficacies of treatments and to study other outcomes among women with mental illnesses. However, achieving recruitment goals has been difficult for researchers. The objective of this study was to examine the success rates of different strategies for recruiting women into clinical trials for the prevention and treatment of postpartum major depression. This is a descriptive study designed to examine which recruitment efforts, over a 4-year period, yielded women who participated in the studies. It was conducted in an outpatient clinic affiliated with a major urban teaching hospital where women of childbearing age sought treatment for affective disorders. All women (n=589) who called about our research studies were systematically screened. The women were either pregnant or in the postpartum period. Our results show that in both studies, obstetricians and other healthcare professionals referred a large number of women. These referrals yielded relatively high rates of participation. However, other methods were differentially successful for the two studies: media appearances and advertising resulted in almost 50% of the screening calls for the prevention study, whereas mass mailings and other promotional materials were more effective for the treatment study. Those materials contributed 44% of screening calls for the treatment study. Our conclusions were that women were most likely to enter the research studies when referred by an obstetrician or other professional. # 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction Before 1977, women of childbearing age were excluded for Phase I & Phase II clinical trials. They could only be included if animal studies determined there were no effects of the outcome on the fetus or female fertility. The paradox is that all women were excluded if they were of childbearing age. Only AIDS and cancer studies included women (Merkatz et al., 1993). In the 1980s, public health leaders drew attention to the inequitable representation of women and minorities in National Institutes of Health-funded studies. The GAO reviewed participation of women and recommended that the FDA needed to ensure gender representation in prescription drug testing. Their study also found that reports from trials that included women did not examine gender differences in outcomes [GAO Pub# GAO/ * Corresponding author. Tel.: +1-215-503-2817. E-mail address: [email protected] (K.S. Peindl).
HRD-93-17, 1993). New guidelines from the FDA in 1993 stated that women should be included in ‘‘reasonable’’ numbers in all clinical trials. The National Institutes of Health (NIH) published An Outreach Notebook (La Rosa et al., 1997) that offers guidelines for the inclusion of women and minorities in research protocols. To date, NIH now mandates these guidelines. A recent article published the results about the representation of women in randomized trials of acute coronary syndromes (Lee et al., 2001). The authors reported that enrollment of women into trials has increased by 25% but this is still below the proportion of women with myocardial infarctions in the general population, which is 43%. As of 2001, women were still under represented in breast cancer trials (Evans et al., 2001). The lifetime risk of breast cancer in the United Kingdom is 40%, but the enrollment into three separate trials for breast cancer has been 10%. Women who were eligible did not enter the trials and 60% chose to be screened. Even with women being underrepresented in
0022-3956/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved. PII: S0022-3956(02)00086-9
118
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
research studies, great strides in understanding women’s health issues have occurred over the past decade. Two large trials, The Women’s Health Initiative (WHI) and The Postmenopausal Estrogen/Progestin Interventions Trial (PEPI) are studies specifically designed to address inequities in women’s health relating to heart disease, cancer, osteoporosis and the effects of hormone replacement on heart disease risk in postmenopausal women. Both trials have had success in recruiting women. Recruitment efforts for the WHI included mass mailings, community outreach, distribution of brochures at many community events, and professional referrals. The enrollment of women into each of these trials was highly successful (Lewis et al., 1998; Johnson et al., 1995). Several investigators have published reports on strategies for recruiting women. Lovato et al. (1997) published a summary of recommendations for the recruitment of women into clinical trials, which included flexible appointment schedules, provisions for daycare, placement of advertisements that target women, and recruitment for the trial within female-dominated practices. Five well-planned strategies were used to recruit women into a 3-year study of caregivers for chronically disabled adults (Patrick et al., 1998). In this trial, the number of participants required to detect differences among the study groups was 800. Implementation of the five strategies allowed for successful recruitment of 844 women. First, the research staff contacted 48 formal service agencies that served women who might be eligible for the study. This strategy accounted for 29.3% of the total sample; 85.4% of those initially recruited completed the study. Second, the research staff presented material about the study to support groups. This strategy accounted for 9.2% of the participants, with a 97.5% retention rate throughout the study. Third, enrolled study participants recommended other women to the researchers, who contacted those potential participants. This strategy accounted for 37.3% of the total sample; their study completion rate was 88.9%. Fourth, the investigators used media and promotional materials (press releases, radio and newspaper advertisements, and distribution of brochures and flyers) to inform potential participants. This method yielded an unknown percentage of the total sample. The fifth and final strategy used a modified demographic sampling unit, which resulted in 0.7% participation and a moderate completion rate of 79.3%. Several investigators have had relative success in recruiting women through mass mailings or face-to-face screening (Lewis et al., 1998; Gilliss et al., 2001; Baines, 1984). In one study, investigators needed subjects for a three-site trial to reduce dietary fat intake in postmenopausal women. By mass mail, they contacted 210,000 women, 11.5% were randomized into the trial (Lewis et al., 1998). The second study was a 5-year investigation of the health of women mid-life through menopause transition (Gilliss et al., 2001). In this study,
face-to-face screening yielded the highest percentage of participation (74%). In the third study, investigators in a Canadian breast cancer screening program had success by following up mass mailings with telephone contacts (Baines, 1984). Of the women contacted by this method, 26.4% entered the study. However, the number of participating women (from mass mailing and phone contact) was less than 1% of the total sample needed for screening. In all three studies, other recruitment methods yielded small numbers of enrolled women. Another group of investigators utilized the strategy of recruiting women from a primary care network into a genetic counseling study for breast cancer risk (Hlemes et al., 2000). Mass mailing used for the recruitment lasted 11 months. Of the women contacted, 8.6% entered the study. The investigators reported that even fewer women wanted to remain in the protocol after learning of future commitments of the research study. Some investigators have examined women’s reasons for agreeing to participate in a research study. Krieger et al. (2001) held focus group sessions for cancer patients who had been asked to participate in a research study. They asked attendees what factors affected their decision to participate. The element with the most positive impact for participants was a pre-enrollment discussion with research staff about the time commitment involved in the study. This paper describes our efforts to recruit women into two clinical trials. The first trial was for the prevention of postpartum major depression and the second trial was for the treatment of postpartum depression. In our studies, we encountered two additional recruitment issues because of our focus on the mental health of pregnant and postpartum women. First, it was necessary to rely on other medical specialists’ practices to refer women to our studies. Other investigators have also described the problem of referrals from other clinical departments (Hlemes et al., 2000). Second, we faced the obstacle of a lack of awareness about mental illness among pregnant and postpartum women. Our two studies were approved by the IRB of the University Hospitals of Cleveland. The first trial was a randomized, double-blind, clinical trial that determined the efficacy of nortriptyline compared to placebo for the prevention of postpartum depression. In this trial, women were healthy during the current pregnancy but had previously experienced an episode of postpartum depression (Wisner et al., 2001). The second trial is a randomized, double-blind, controlled clinical trial comparing nortriptyline versus sertraline for the treatment of postpartum depression.
2. Method We screened all calls to our women’s mental health clinic, which specializes in the treatment of affective
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
disorders among women of childbearing age. We recorded information on a standardized screening form (see Appendix). The form included questions about pregnancy or postpartum status, demographic information, insurance coverage, and original source of information about the study. When the caller was referred by a healthcare professional, we recorded the professional’s name, phone number, and specialty (e.g. obstetrician, mental health therapist, midwife). Other information sources included promotional materials, media appearances, and advertisements. 2.1. Treatment study A mailing was sent by first-class mail to all women who delivered in the previous month at the University Hospitals of Cleveland. On average, 450 infants were delivered each month. The mailing included a letter of congratulations on the birth of a baby, information about postpartum depression and our clinic, our expression of concern for their postpartum well-being, and a flyer about our postpartum treatment study. Both the letter and flyer were approved by the IRB. Two pamphlets were also sent: Mom never had a chance to be involved in medical research, published by the Society for Women’s Health Research (2000), and Volunteering for a clinical trial, published by CenterWatch (1999). We enclosed a self-rating scale for postpartum depression; the Edinburgh Postnatal Depression Scale (EPDS, Cox et al., 1987), with instructions to call us if they scored 12 or higher, felt depressed, or had questions. We encountered no difficulty in women completing the screen. They were willing to complete it as a check of their emotional wellness. We obtained the names and mailing addresses of the women from the hospital database after IRB and hospital approval. The mailing list was also available to other companies who direct mail to mothers with newborns. The women in the mailing did not have, to our knowledge, any mental illness but some were at risk for having an onset of major depression. We explained that risk in the enclosed letter. Today, it would be unlikely that an IRB would approve this type of mailing because of the breach in confidentiality of these patients. Study brochures were sent to all obstetricians in the Cleveland area, with a request to place them in their offices for patients. The research staff also visited offices and requested permission to place flyers and brochures in the waiting rooms. We found that the research staff placing study information in the offices was most successful. With the obstetricians’ packet of brochures, we included information about our clinical practice and research studies. In an enclosed letter to the doctors, we asked to schedule a meeting to present information about the clinical practice and our studies. We presented to three out of ten large obstetrical practices that were willing to meet with us. We also contacted 150
119
obstetrics and gynecology physicians by mail. Twentyfour (16%) of the contacted physicians referred to our studies. The Principal Investigator (KLW) reached other obstetricians by presenting our research projects at meetings of the local Obstetrics and gynecology Society. Several members of the research staff gave talks to family medicine residents, nurse midwives, psychiatric and obstetric residents, and the nursing department within the university. At each presentation, we provided flyers and brochures about our studies and our clinical services. In addition, the research staff placed brochures and flyers in libraries, hospitals, and clinical practice buildings throughout the city. We also utilized the media for recruitment purposes. The PI (KLW) and several clinicians served as guests on local television talk shows about postpartum depression and other mental illnesses that are more prevalent among women. We placed advertisements about the studies and our clinic in newspapers, on the radio and the scroll bar of a local television weather station. 2.2. Prevention study Our recruitment strategies for the prevention study were similar to those for the treatment study. A mass mailing was sent in a pre-admission packet to all women who delivered at University Hospital. We did not have access to the names and addresses for this mailing that was done by hospital personnel. It included a letter, a brochure about the study, and an information sheet on postpartum depression. These enclosures were approved by the IRB of the hospital. Brochures, posters with informational tear-offs, and flyers were placed in obstetrics offices throughout the city. We presented to the local Obstetrics and gynecology Society, midwives, nurses, and obstetrics and gynecology offices. Mass media presentations included morning television talk shows; a segment on PrimeTime Live (ABC); and cable television, radio, and newspaper advertisements. For the prevention study, we also utilized a face-toface screening process. One of our research staff (nurse) gained access to a large hospital-based obstetrics clinic and screened pregnant women for history of postpartum major depression. The nurse was time-limited and her clinical obligations permitted only 1 day per week to screen. Because of the strict eligibility requirements for the study, we did not find any eligible women. If we would have had a full-time staff member who was able to screen daily, we may have recruited women for the study.
3. Results Table 1 shows the number of contacts generated by each type of information source for each study. Professional
120
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
Table 1 Number of women screened, by information source and study Information source
Number of contacts per study (%)
Table 2 Number of women participating in treatment study, by information source Information source
Number screened Total=283
Number participating (% of total participants) N=77
(48) (28) (7) (4) (3) (4) (2)
Professional referrals—total Obstetrician Nurse Counselor/social worker Pediatrician Psychiatrist Midwife
136 79 19 11 9 11 7
38 21 7 2 5 2 1
(49) (27) (9) (3) (6) (3) (1)
16 5 11
(6) (2) (4)
Other referrals—total Insurance advocate Family member or friend
16 5 11
8 4 4
(10) (5) (5)
(6) (3) (3) (0) (0)
125 103 19 1 2
(44) (36) (7) ( <1) ( <1)
Promotional materials—total Mass mailing (postpartum) Brochures/flyers Hospital video Website
125 103 19 1 2
30 25 5 0 0
39 (32) (6) (0) (0)
64 45 11 8
(21) (15) (4) (3)
2 2 N/A 0
( <1) ( <1)
Local television talk shows Radio advertisements
2 4
0 1
(0) (1)
Media advertisements—total Cable television Radio Newspaper
88 65 2 21
(29) (21) ( <1) (7)
4 0 4 0
(1) (0) (1) (0)
Face-to-face screening Ohio Task Force
44 1
(14) ( <1)
N/A 0
(0)
Prevention N=306
Treatment N=283
Professional referrals —total Obstetrician Nurse Counselor/social worker Pediatrician Psychiatrist Midwife
83 58 5 2 0 12 6
(27) (19) (2) ( <1) (0) (4) (2)
136 79 19 11 9 11 7
Other referrals—total Insurance advocate Family member or friend
8 0 8
(3) (0) (3)
Promotional materials—total Mass mailing Brochures Hospital video Website
18 8 10 0 0
Media appearances/news—total Local television talk shows PrimeTime Live (ABC) Other television news
(0)
referrals resulted in a high number of completed screens for both studies. In the treatment study, the mass mailing was particularly effective in generating contacts, whereas in the prevention study, the cable television advertisements resulted in more contacts. 3.1. Treatment study We screened 283 women for the postpartum depression treatment study over a 4-year period. Thirty-five percent (n=99) of the women refused the study after the initial screen. Among those who refused, a physician referred 28% of the women and 38% received the postpartum letter. Women had an initial evaluation to determine eligibility. Among those women (n=107), 38 percent of the women were ineligible. Seventy-seven women (13%) agreed to participate in the study. One woman refused medication after randomization into the trial. Table 2 shows the number of actual study participants generated by each type of information source for this study. The greatest proportion of women who contacted
us (48%) were referred by healthcare professionals—a group comprised of obstetricians, nurses, counselors, social workers, pediatricians, psychiatrists, and midwives. Women referred by these professionals made up almost half of our study participants. The second largest number of phone contacts to the clinic came from the mass mailing, making up 36% of the women screened and ultimately 32% of our study participants. Interestingly, although the number of screening contacts generated by insurance advocates was small (n=5), a high percentage of these women (80%) entered the study. The insurance companies had no stake in the research. All visits and medications were free to the participant, but two women within the insurance companies were advocates for the study and made referrals to trial. The staff salaries and medications were furnished from grant expenses. 3.2. Prevention study We screened 306 women for the postpartum depression prevention study over a 3.5-year period. Forty percent of the women screened did not fit criteria for entry into the trial and 20% of the women refused participation after the phone screen. Among the screened women, 121 were eligible for the trial and fifty-seven women entered the study. Five women declined medication after randomization and one woman became manic during the first week postpartum. For this study, the participation rate was 42% among eligible women. Referrals from healthcare professionals again yielded high screening and participation rates. However, the
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
results for other sources were slightly different from those for the treatment study. Table 3 shows the number of study participants generated by type of information source for this study. Media appearances and advertisements generated a large volume of screening calls. However, they did not always convert to high rates of study participation. Appearances on local television talk shows resulted in 45 screening calls, but these women ultimately comprised only 7% of the sample. Advertisements on cable television were more successful, resulting in 65 screening calls and 21% of study participants. Other recruitment strategies yielded very small numbers of women entering the study; notably, the face-to-face screening process generated no study participants.
4. Discussion The limitation of this study is that the two trials were specifically designed for treatment and prevention of major depression in postpartum women. We needed to be sensitive to mental health issues among the physicians and the women that we were seeking to enter the trials. The trials were conducted in an academic setting, but we recruited women throughout the metropolitan Table 3 Number of women participating in prevention study, by information source Information source
Number screened Total=306
Number participating (% of total participants) N=57
Professional referrals —total Obstetrician Nurse Counselor/social worker Psychiatrist Midwife Family member/friend referrals
83 58 5 2 12 6 8
25 20 2 0 1 2 2
(44) (35) (4) (0) (1) (4) (4)
Promotional materials—total Mass mailing (pre-admission) Support group mailing Brochures/posters
25 8 7 10
9 5 0 4
(16) (9) (0) (7)
Media appearances/news—total Local television talk shows PrimeTime Live (ABC) Other television news
64 45 11 8
5 4 1 0
(9) (7) (2) (0)
Media advertisements—total Cable television Radio Newspaper
82 65 2 15
16 12 0 4
(28) (21) (0) (7)
Face-to-face screening Ohio Task Force
44 1
0 0
(0) (0)
121
community. Recruitment for both studies depended largely upon referrals from physicians and other professionals outside the academic community. At first, referrals from physicians were difficult. We found that at least one year was necessary to develop relationships within the community and to establish a referral network of obstetricians. When developing these relationships, we emphasized our interest in their patients’ wellbeing rather than placing emphasis on mental illness. It took time to build their trust and we established trust with successful treatment of their patients or having one of their patient (by self-referral] enter a study and successfully complete it. Our report shows that the relationship that the patient has already established with her healthcare provider encourages research participation when a referral is made. For successful recruitment, strategies should be planned and implemented before beginning the study, with follow-up plans formulated to keep the studies fresh in the minds of health professionals and to replenish brochures and flyers on a monthly basis. During the planning stage, the research team should realistically address the labor and time commitments of the proposed methods of recruitment. For example, mass mailings for 450 deliveries per month take 4 h to complete. Monthly flyer development and copying of enclosures in the mailing takes 6 h. Delivering flyers and brochures to obstetric and gynecology offices takes twice a month with deliveries made twice a week. These commitments may also include after-office-hour presentations for television and radio shows. Media outlets are attractive for recruitment purposes because of the wide audience they reach. However, we had mixed results with our media-related strategies. Although they generated a high number of calls, they yielded low rates of participation. Media advertisements can be costly, so it is wise to plan carefully the timing and placement of any ads. For example, we found that our advertisements on the cable television weather channel (for the prevention study) were successful only during the winter months, when people were more concerned with weather changes. We had no response to this recruitment strategy during the summer. Referrals by health care professionals were our best source of potential study candidates (49, 44% participation rates). Although time consuming, mass mailings gave us 32% of the sample in the treatment study. Careful use of TV (in our case, cable weather channel) did give us participants for the prevention study (21%). We cannot measure the impact of the educational programs presented on TV or radio talk shows about the importance of women’s mental health. Any information given to the public helps women come for treatment as in the case of breast cancer awareness. However, more information needs to be disseminated about the importance of research to the knowledge and treatment of
122
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
psychiatric disorders among women. Without their trust in the research team and in the research endeavor, we will still encounter difficulties in recruiting women into clinical trials. We make the following suggestions for recruitment of women into mental health research: 1. Establish relationships with psychiatric and primary care physicians, and obstetricians. Depending on the type of study, different practices can be targeted for education about the research and women’s mental health. In our studies, we contacted obstetrics and gynecology practices, midwives, pediatricians, obstetrical nurses, and social workers about the importance of the studies to the treatment and prevention of postpartum depression. Family Practice physicians and insurance companies are another potential source for referrals. 2. Encourage primary care physicians to be ambassadors for the project by educating them about mental illness and the importance of your research to the well-being of their patients. Provision of information to physicians about their patients’ progress in research protocols is imperative to establish trust. 3. Inform the research staff of the time commitments involved in developing and implementing
recruitment strategies. This time commitment includes developing strategies through regular weekly meetings, calling offices and making appointments to meet with specialists and their staff, arranging to deliver flyers and brochures to specialists’ offices, creating flyers for mass mailings, obtaining information (addresses) for mass mailings, and presenting informative talks to departments, societies, and public health officials about research and women’s mental health issues. These presentations may be after regular office hours. 4. Educate the public—which includes potential research participants—not only about the mental health issue being studied, but also about the overall importance of research in improving mental health care. The information can be disseminated by various methods; for example, we used mass mailings to provide women with information about postpartum depression and participating in clinical trials.
Acknowledgements Funding for the project came from National Institutes of Mental Health MH53735 and MH57102.
Appendix follows on next page.
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
Appendix
123
124
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125
References Baines CJ. Impediments to recruitment in the Canadian national breast cancer screening study: response and resolution. Controlled Clin Trials 1984;5:129–40. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry 1987;150:782–6. Evans DGR, Lalloo F, Shenton A, Boggis C, Howell A. Uptake of screening and prevention in women at high risk of breast cancer. Lancet 2001;358:889. Gilliss CL, Lee KA, Gutierrez Y, Taylor D, Beyene Y, Neuhaus J, et al. Recruitment and retention of healthy minority women into community- based longitudinal research. J Women’s Health & Gender-based Medicine 2001;10:77–85. Hlemes AW, Bowen DJ, Bowden R, Bengel J. Predictors of participation in a primary care physician network. Cancer Epidemiology 2000;9:1377–9. Johnson S, Mebane-Sims I, Hogan PE, Stoy DB. Recrutiment of
postmenopausal women in the PEPI trial. Postmenopausal Estrogen/Progestin Interventions. Controlled Clin Trials 1995;16(4):20S– 35S. Krieger N, Ashbury F, Cotterchio M, Macey J. A qualitative study of subject recruitment for familial cancer research. Ann Epidemiology 2001;11:219–24. La Rosa JH, Parron D, Alvarado M. Outreach notebook for the NIH guidelines on inclusion of women and minorities as subjects in clinical research. National Institutes of Health publication No. 97-4160; 1997. Lee PY, Alexander KP, Hammill BG, Pasquali SK, Peterson ED. Representation of elderly persons and women in published randomized trials of acute coronary syndromes. JAMA 2001;286:708–13. Lewis CE, George V, Fouad M, Porter V, Bowen D, Urban N. Recruitment strategies in the Women’s Health Trial: feasibility study in minority populations. Controlled Clin Trials 1998;19:461–76. Lovato LC, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruitment for controlled clinical trials: literature summary and annotated bibliography. Controlled Clin Trials 1997;18:328–57.
K.S. Peindl, K.L. Wisner / Journal of Psychiatric Research 37 (2003) 117–125 Merkatz RB, Temple R, Sobel S, Fieden K, Kessler DA. Women in clinical trials of new drugs—a change in Food and Drug Administration Policy. NEJM 1993;329:292–6. Mom never had the chance to be involved in medical research [pamphlet]. Society for Women’s Health Research; Washington, DC; 2000. Patrick JH, Pruchno RA, Rose MS. Recruiting research participants:
125
a comparison of the costs and effectiveness of five recruitment strategies. The Gerontologist 1998;38:295–302. Volunteering for a clinical trial [pamphlet]. Boston, MA: CenterWatch, 1999. Wisner KL, Perel JM, Peindl KS, Hanusa BH, Findling RL, Rapport D. Prevention of recurrent postpartum depression: a randomized clinical trial. J Clin Psychiatry 2001;62:82–6.