Contraception 85 (2012) 363 – 368
Original research article
Bridge over troubled waters: considerations in transitioning emergency contraceptive users to hormonal methods Dawn Chin-Quee a,⁎, Laura Hinson b , Kelly Ladin L'Engle a , Conrad Otterness a , Barbara Janowitz a a FHI, 2224 E NC Highway 54, Durham, NC 27713, USA Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205-2103, USA Received 9 May 2011; revised 28 July 2011; accepted 29 July 2011
Abstract Background: As emergency contraceptive pills (ECPs) become increasingly available through pharmacies, concerns about potential overuse of this product have emerged. In response, bridging women from ECPs to ongoing contraception was advanced as a solution. Study Design: We collected information in Ghanaian pharmacies on ECP users' sexual activity, use of contraceptive methods and reasons for buying ECPs. Further, two behavioral indicators were examined to determine whether a woman should consider using an ongoing contraceptive method: how often she has sex and how she uses ECPs. Results: Of the four types of ECP users, stratified by those two indicators, only women who have sex frequently and use ECPs as their main contraceptive method would be appropriate for, but not necessarily amenable to, bridging. Conclusions: The challenges of bridging to meet the contraceptive needs of women are discussed in light of the characteristics of emergency contraceptive users and suggest that bridging is not as straightforward as initially conceived. © 2012 Elsevier Inc. All rights reserved. Keywords: Emergency contraception; Contraceptive bridging; Hormonal contraception; Oral contraceptive pills; Pharmacies; Ghana
1. Introduction Emergency contraceptive (EC) pills (ECPs) are available for women who have had unprotected sex — either due to method failure (condom breakage, slippage), incorrect use [late injections, missed oral contraceptive pills (OCPs)], nonuse of a contraceptive method or as a result of sexual assault. Originally offered by prescription only, ECPs are now provided behind the counter (i.e., with pharmacist counseling) or over the counter in many countries. This expedites the acquisition process and improves the efficacy of the method by shortening the interval between the unprotected sex act and ingestion of the tablet(s) [1,2]. The number of dedicated brands of ECP products and their sales has increased following the shift from prescription to nonprescription status, allowing greater access to the method (E. Westley, personal communication, September 23, 2009). ⁎ Corresponding author. Tel.: +1 919 544 7040, 919 493 1950; fax: +1 919 544 7261. E-mail address: email@example.com (D. Chin-Quee). 0010-7824/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.contraception.2011.07.014
With this increased access, anecdotes of misuse and abuse of ECPs in the private sector have emerged [3,4]. Media reports suggest that women are making repeat purchases of ECPs in pharmacies [5–7]. However, this claim has not been substantiated by well-designed, population-based research. In fact, many in the EC community support repeat use of ECPs, since there is currently no evidence that the practice is harmful  and it offers every woman a second chance to prevent an unwanted pregnancy [9,10]. Initially, the ECP community responded to the prospect of regular repeat use of this method with interventions to transition, or “bridge,” ECP users to longer-term, ongoing methods 1 such as OCPs, injectables like Depot medroxyprogesterone acetate and the intrauterine device (IUD) [11,12]. However, both EC advocates and researchers did not fully appreciate that ECP users were not a homogeneous group, as bridging interventions did not draw distinctions 1
Throughout this article, methods referred to and contrasted with ECPs as longer-term, ongoing or more effective methods of contraception include OCPs, injectables, implants and the IUD.
D. Chin-Quee et al. / Contraception 85 (2012) 363–368
between ECP clients' current contraceptive practices and their reasons for purchasing ECPs. This might explain the dearth of success stories — published or unpublished — of bridging pharmacy-based ECP users to more effective contraceptive methods. For instance, while some women use ECPs when their contraceptive method has failed, others use them after sporadic, unprotected and/or unplanned sex and may not need or desire a longer-term contraceptive method. In Jamaica, this may have been the case where over 70% of pharmacy-based ECP users followed up over a 6month period reported that they were not interested in adopting regular OCPs — even when offered an incentive to do so . A study conducted in Nairobi, Kenya, also pointed to the challenges of bridging in a pharmacy setting, as ECP purchasers provided with free information on effective and ongoing contraceptive methods reported the gesture as intrusive  and were described by pharmacy staff as uncooperative after being given the information . In both of these studies, women's current contraceptive profile and sexual activity were not taken into account in the design of their respective bridging interventions. Frequency of sexual activity and whether or not ECPs are used as a main or backup method may affect whether a woman would be amenable to bridging, as these factors are expected to affect both use of other methods and frequency of ECP use. These and similar indicators have been reported in previous studies [16,17], but they have not been used to explore the potential for bridging. An understanding of how these behaviors may affect use of ECPs and other methods is an important step in determining how acceptable and effective a bridging intervention might be for ECP users. As part of a formative assessment to design an intervention to bridge women from ECPs to ongoing methods of contraception, we collected information in Ghanaian pharmacies on ECP users' sexual activity, their use of contraceptive methods and factors precipitating their decision to buy ECPs. We use the findings from this assessment to illustrate important considerations in the design of bridging interventions and, in doing so, highlight the characteristics of these ECP users in Ghana that suggest alternative approaches in meeting the reproductive needs of similar women.
that reported the highest volume of sales were asked to participate in the study. Pharmacy owners at the 12 sites agreed to allow trained interviewers to be posted at their establishments for a period of 12 nonconsecutive business days, during times that coincided with their highest sales of ECP products. Pharmacy staff assisted the interviewers by informing ECP purchasers that a study on emergency contraception was being conducted on site. If the ECP client was interested, he or she was directed to speak with the interviewer. Interviewers recorded the gender, age and intended user of the ECP product for each client they intercepted. However, only females 18 years and older who were buying ECPs for their own use were eligible to be interviewed after informed consent was obtained. Of the 329 individuals intercepted (including 61 males and 13 minors), 232 eligible women were interviewed. In addition to demographic information, women were asked about their ECP use, other contraceptive methods used, their reasons for buying ECPs that day, whether they considered ECPs to be their main method of contraception and frequency of sexual activity in a typical week. We categorized women into four groups based on two factors: their sexual activity and whether they considered ECPs to be their main contraceptive method. Frequent sexual activity was defined as having sex at least twice a week and was distinguished from sexual activity once or less per week, which is the presumptive definition of infrequent sex used by the World Health Organization and others . Women who reported ECPs as their main method may have also used other methods of contraception and were contrasted with women who reported greater reliance on another or a combination of other methods besides ECPs for contraception. These four groups were also profiled by their method mix and the frequency of their ECP purchases in the last 12 months. The study protocol and data collection materials were approved by Family Health International's Protection of Human Subjects Committee and the Ghana Health Service's Ethics Review Committee.
3. Results 3.1. Characteristics of ECP users
2. Methods With assistance from the Ghana Social Marketing Foundation, we obtained a list of 55 pharmacies in the Greater Accra area that sold dedicated ECP products. 2 The 12 pharmacies 2 Not all products sold to pharmacy clients as emergency contraception were the dedicated products (Postinor 2®, Norlevo® consisting of 2 tablets each containing 0.75 mg of the active ingredient levonorgestrel) on which we based our sale volume criteria for site selection. About 13% of our sample bought Primolut N, a product sold as 5-mg tablets of norethisterone that is indicated for the postponement of menstruation or the treatment of endometriosis, menorrhagia and dysmenorrhea. Its off-label use as emergency contraception is common in Ghana.
Overall, ECP users were young and well educated (Table 1). About half were under 25 years, and a high proportion was still in school. More than half were single, and most did not have children. The majority were gainfully employed. Approximately one third considered ECPs to be their main method of contraception, and about two thirds had sex twice or more per week. 3.2. ECP users categorized by behavioral characteristics Classification by frequency of sexual activity and whether ECPs were considered the main method of contraception yielded the following groups (Table 2):
D. Chin-Quee et al. / Contraception 85 (2012) 363–368
against sexually transmitted infections (STIs) and not for contraception. Group 2: ECP is not the main method; high sexual activity. Women who had sex frequently and did not consider EC to be their main method constituted a plurality of 107 women, or 46% of our sample. Almost all reported using a contraceptive method, and like their counterparts in Group 1, this method was primarily condoms — alone or in combination with other methods. As in Group 1, the small number of women who reported using no contraceptive method at all also reported infrequent condom use. Group 3: ECP is the main method; low sexual activity. Only 38 women (16%) were in this group. As would be expected, use of a method other than ECPs was much lower in this than in the previous two groups. Less than half used a contraceptive method other than ECPs, but of those who did, condoms alone or in combination with other methods was again the most prevalent. Group 4: ECP is the main method; high sexual activity. Just over 40% of the sample, or 18 women in this group, used a method (primarily condoms) other than ECPs. Women in this group were more likely than in other groups to purchase ECPs four or more times in the previous 12-month period. This finding is expected, as these women have frequent sexual activity but low reliance on methods other than ECPs.
Table 1 Characteristics of ECP users in Ghana Client characteristics
Age distribution (years) 18–24 25–34 35+ No. of children 0 1–2 3 or more Highest level of education completed Primary or less Middle or secondary Higher School attendance: currently in school Work status: works for money Marital status Single In union/married Divorced/widowed/separated Use of ECPs: considers ECPs the main method Sexual activity Has sex once per week Has sex 2–3 times per week Has sex 4 or more times per week Does not have sex regularly on a weekly basis
48 47 5 61 33 6 3 50 47 44 68 57 38 5 35 18 45 20 18
Group 1: ECP is not the main method; low sexual activity. Forty-four women, or 19% of the sample, fell into this group. Of these, 84% used a contraceptive method other than ECPs, with 68% using condoms alone or in combination with other methods (usually traditional methods). The 16% who stated that they did not use contraception reported infrequent condom use, suggesting that they used condoms for protection
3.3. Reasons for buying ECPs One hundred thirty-seven women (59%) bought ECPs on the day they were interviewed because they had used no contraceptive method at all during their last sex act. Method
Table 2 Percent using contraception, contraceptive methods used and frequency of ECP purchase in the last year by frequency of sex and whether ECP is the main contraceptive method
Contraceptive use a Uses a method other than ECPs Condom only Condom and other methods Hormonals Traditional method b Not specified/other Not using a method
Total (n=232), %
ECP not the main method
ECP as the main method
Group 1, sex once or less per week (n=44), %
Group 2, sex two or more times per week (n=107), %
Group 3, sex once or less per week (n=38), %
Group 4, sex two or more times per week (n=43), %
75 24 32 10 7 1 25
84 39 29 14 2 0 16 c
95 23 45 14 12 1 5c
43 24 13 0 3 3 57
42 9 21 7 5 0 58
27 59 11 2
15 64 18 4
24 55 11 11
2 51 23 23
Frequency of ECP purchase last year a 1 time 16 2–3 times 59 4–6 times 16 7 or more times 8 a b c
The percentages may not add to 100 due to rounding. Traditional methods include withdrawal, safe days, calendar, herbs, etc. These women do not use ongoing methods but have reported using condoms infrequently.
D. Chin-Quee et al. / Contraception 85 (2012) 363–368
that these women do not need a more effective method like OCPs to prevent unwanted pregnancy. However, an additional consideration of bridging to a more effective, ongoing method is that it may lead to a lower use of condoms. Our sample of ECP users is comparable to others in that they were also largely self-identified as condom users [19–22]. Thus, encouraging this group to bridge may give rise to a decrease in condom use that, in turn, could result in an increased risk for HIV and other STIs. Like ECPs, these methods offer no protection against HIV and STIs, making the identification of appropriate candidates for bridging even more challenging and calling into question the “valueadded” of such an approach in places where there are high HIV rates. A better strategy may be to assist these women in becoming better condom users, a difficult but not impossible task [23,24], while encouraging them to continue using ECPs as needed. This would simultaneously meet the pregnancy and STI prevention needs of this larger group. By contrast, approximately one third of our sample regarded ECPs as their main method, though 40% of these women also used another contraceptive method. Thus, about 60% of women in Groups 3 and 4 are protected from pregnancy solely through ECP use. Women who have more frequent sex (Group 4) face a high cumulative risk of pregnancy — especially if they have unprotected sex multiple times within the same menstrual cycle and rely solely on ECPs for protection. Emergency contraception as a main method may not be ideal for this group. However, other women use ECPs as a main method but have sex infrequently (Group 3). If shown to be effective when used routinely, EC may be ideal for their contraceptive needs. Our study results not only identify who may or may not be appropriate for bridging but, in accounting for the user characteristics and behaviors presented here, also question the need (and our ability) to bridge all repeat users of ECPs. In the private sector where most ECPs are obtained, bridging
failure accounted for reasons given by 32% of the sample or 74 women who bought ECPs that day. A small proportion (10%) of women also bought the method in advance of need. Because many of these women were condom users, data are also presented on frequency of condom use. Of the women who considered ECPs to be their main method of contraception, none used condoms all the time in the last 3 months; only 8% of their counterparts — self-identified, regular condom users — reported the same (Table 3). 4. Discussion 4.1. Who would be good candidates for bridging? To illustrate considerations that should be made in determining who would be appropriate for bridging, we draw on our previously described four groups of women. We also focus on the scenario of bridging from ECPs to OCPs, because in most cases, OCPs are sold — without a prescription — in the same pharmacies as ECPs. Thus, with a few exceptions (such as the United States), ECPs are no more accessible than OCPs (or vice versa). Given the importance of convenience and access, a bridging intervention that encourages the transition to ongoing OCP use (as opposed to injectables in a clinic setting) would hold the most promise. Most (65%) of the women in our sample did not consider ECPs to be their main method. On the surface, they would seem not to need bridging because ECPs are unlikely to be used by them repeatedly. However, this sample of Ghanaian women also reported very low use of more effective methods such as pills and injectables, with greater reliance on condoms alone and in combination with mostly traditional methods. Further, since over 70% of them (Group 2) have sex twice or more per week and because their condom use is inconsistent at best, it would be hasty to draw the conclusion
Table 3 Women who use condoms for contraception (or STI prevention) by reasons for purchase of ECPs and whether or not ECP is considered the main contraceptive method Condom users
Does not use condoms for contraception (n=102), %
Total (n=232), %
50 36 7 6 1
71 1a 16 12 5
59 19 10 10 3
8 38 43 10 2
1a 0 8 28 63
4 20 28 19 28
ECP: main method (n=27), %
ECP: not the main method (n=103), %
Reasons for purchase of ECPs Used no contraceptive method Condom slipped/broke/tore Missed OCPs Bought ECPs in advance in case needed later Late for injection
52 22 4 22 0
Frequency of condom use in the past 3 months All the time Most of the time Some of the time Rarely Never
0 30 48 22 0
Since this respondent did not report using any method of contraception, she very likely considered her condom use as STI prevention only.
D. Chin-Quee et al. / Contraception 85 (2012) 363–368
has yet to be proven achievable. It follows on the failure to demonstrate that increased access to ECPs results in decreased rates of unintended pregnancies and abortions . The jury is still out on bridging, but the notion of recommended repeat use of ECPs has been gaining strength among advocates of increased access to EC [26,27]. Since experience with pill regimens of similar or higher dosages than current progestin-only ECP formulations “suggests that the likelihood of serious harm from at least moderate repeat use is low” , many in the family planning community believe that it is better to use ECPs on demand rather than nothing at all. In response, a third course of action, the possible development of a coitally dependent hormonal method, is currently being explored. 4.2. Women's reasons for buying ECPs The reasons for buying ECPs add yet another layer to considerations in designing bridging interventions and get at the heart of why women may prefer to use ECPs as needed rather than adopt a more effective hormonal method. What may be more important to a woman are the convenience and ease of obtaining ECPs at the pharmacy (relative to methods like prescription OCPs, injectables and the IUD) and the ability to address her risk of pregnancy only when she perceives it to be immediate. What also appears to be less important is the cost of ECPs. In absolute terms, the cost of ECPs in most developing countries is not prohibitive (as compared to the United States and UK, for example). Nevertheless, even in pharmacy settings, a single-use package of ECPs costs four to five times more than a 1month supply of nonpremium brands of OCPs. 3 Based on sales in affluent, middle- and low-income areas in countries such as Ghana, Jamaica and Kenya, the cost of ECPs does not appear to be a factor in women's decisions to use this method. Indeed, message testing that was part of our formative research in Ghana indicated that the lower cost of OCPs was not incentive enough to switch from ECPs. 4.3. Limitations This was a small study of women who purchased ECPs in selected pharmacies in Accra and is not meant to be representative of ECP users in sub-Saharan Africa, Ghana or even Accra. The fact that our study was conducted with a convenience sample increased the likelihood of frequency bias; that is, the probability that we intercepted and interviewed a greater number of frequent as opposed to infrequent users of ECPs. However, our purpose was not to document frequency of ECP use, but to understand the reasons for ECP use and their implications for contraceptive use more generally. Accordingly, the principal behavioral characteristics presented here are universal and, along with 3 Premium combined oral contraceptive brands such as Dian-35 and Jasmine that sometimes require a doctor's prescription can be more expensive than ECPs, as they are marketed to address noncontraceptive benefits such as the control of acne.
the idiosyncrasies of this sample, serve to provide insight into factors that should be taken into consideration in the design of bridging interventions and other efforts to assist women with their reproductive health needs. Specific to this sample, our estimate of the frequency of ECP purchases may be too low, as women may have underreported these purchases. Like self-reported condom use, reports of ECP use may be affected by a social desirability bias  or by the wording of the questionnaire item. However, the highest use was reported in the group with high sexual activity and ECPs as the main method, suggesting that there was internal consistency in the reporting of ECP purchases. Lastly, the reasons for buying ECPs reported in our study were captured by close-ended options, precluding in-depth responses. One also needs to know why women did not use a method during their last sex act, what contributed to the failure and inconsistent use of methods other than ECPs and what drives the decision to purchase ECPs in advance of need, for example. Qualitative research can address this gap in knowledge and should be considered in conjunction with survey research as a mixed-methods approach.
5. Conclusions This study focused on bridging pharmacy-based repeat ECP users as a possible strategy to increase protection against the risk of pregnancy and did so via an examination of users' characteristics and behaviors. Viewed through this prism, the challenges of bridging loom large when confronted with identification of appropriate candidates, possible lack of privacy in these facilities and the inability or unwillingness of pharmacists to counsel ECP clients. Furthermore, as demonstrated by Liambila et al. , clients may not give even a willing pharmacist the opportunity to speak with them about contraceptive options. Based on our findings, it would appear that only the small minority of women who have frequent sex and do not use an ongoing method would be appropriate for and benefit from bridging. Thus, the cost and effort of designing bridging interventions may not be deemed worthwhile. Moreover, if these women “love their ECPs” as found in a small qualitative study of women in Accra , then successful interventions would have to address their reasons for preferring to use ECPs. While the characteristics of ECP users in this study are similar to others, our findings — based on a small sample of Ghanaian women — are speculative. For this reason, additional research is needed to verify and more thoroughly explore the potential for bridging interventions. However, we should not only focus solely on bridging but also encourage testing other strategies such as copackaging or copurchasing of condoms and ECPs. While both of these strategies may discourage condom use and increase
D. Chin-Quee et al. / Contraception 85 (2012) 363–368
exposure to STIs/HIV and possibly also to pregnancy, at the same time, both strategies provide additional choices to reduce risk. Women (and men) may not make the best choices from a public health perspective, but they should be given the tools to broaden their effective options so that they can make appropriate personal decisions concerning risk reduction. Acknowledgments The work on which this study is based was funded by the William and Flora Hewlett Foundation. The authors would like to thank Mr. Alex Banful, Managing Director of the Ghana Social Marketing Foundation, and Dr. Gloria Asare, Director of the Family Health Division, Ghana Health Service, for their guidance and support throughout study development and implementation. They also would like to thank the staff of Research International in Accra for their expert handling of data collection and management. The authors acknowledge Drs. Willard Cates, Elizabeth Raymond, Johannes van Dam and John Stanback for their thoughtful reviews on earlier versions of this manuscript. References  World Health Organization. Selected practice recommendations for contraceptive use. 2nd ed. WHO: Geneva; 2005.  Westley E. Regimen update from the International Consortium for Emergency Contraception. New York: International Consortium for Emergency Contraception; 2003.  Wesangula D. Scandal of birth pills for teenagers. Nairobi: Daily Nation; 2008. Available from: http://www.nation.co.ke/News/-/1056/ 467942/-/tkbtwv/-/index.html.  Blair L. “Morning after” scare: Postinor 2 abuse worries pharmacists. Kingston: Jamaica Gleaner Online; 2003. Available from: http://www. jamaica-gleaner.com/gleaner/20031130/lead/lead1.html.  Yam EA, Gordon-Strachan G, McIntyre G, et al. Jamaican and Barbadian health care providers' knowledge, attitudes and practices regarding emergency contraceptive pills. Int Fam Plann Perspect 2007;33:160–7.  Mawathe A. Kenya concern over pill-popping. Nairobi; 2009. Available from: http://news.bbc.co.uk/2/hi/8145418.stm.  Rao K. Indian docs worry emergency contraception misused. Mumbai: Women's e-News; 2008. Available from: http://www.womensenews. org/story/the-world/081016/indian-docs-worry-emergencycontraception-misused.  Halpern V, Raymond EG, Lopez LM. Repeated use of pre- and postcoital hormonal contraception for prevention of pregnancy. Cochrane Database Syst Rev 2010:CD007595.  Coeytaux F, Wells ES, Westley E. Emergency contraception: have we come full circle? Contraception 2009;80:1–3.
 Westley E. Emergency contraception: a global overview. J Am Med Wom Assoc 1998;53:215–8, 37.  Onyango F, Keesbury J, Faye Y. EC as a “bridge”. EC Afrique Bulletin 2008;5(1):1–2.  Population Council. Mainstreaming EC in Africa: the EC Facilitative Re-granting Program Final Report. Nairobi: Population Council; 2009.  Chin-Quee D, Wedderburn C, Otterness C, et al. Bridging emergency contraceptive pill users to regular contraception: results from a randomized trial in Jamaica. Contraception 2010;81:133–9.  Wanja J. Young Kenyan women top contraception user-study. Nairobi: Daily Nation. June 2, 2010 ed; 2010. Available from: http://www. nation.co.ke/News/Contraception%20high%20among%20women %20in%20mid%2020s/-/1056/930606/-/5an94h/-/index.html.  Liambila W, Obare F, Keesbury J. Can private pharmacy providers offer comprehensive reproductive health services to users of emergency contraceptives? Evidence from Nairobi, Kenya. Patient Educ Couns 2010;81:368–73.  Bastianelli C, Farris M, Benagiano G. Reasons for requesting emergency contraception: a survey of 506 Italian women. Eur J Contracept Reprod Health Care 2005;10:157–63.  Sander PM, Raymond EG, Weaver MA. Emergency contraceptive use as a marker of future risky sex, pregnancy, and sexually transmitted infection. Am J Obstet Gynecol 2009;201:146.e1–6.  UNDP, UNFPA, WHO, et al. Efficacy and side effects of immediate postcoital levonorgestrel used repeatedly for contraception. Contraception 2000;61:303–8.  Blanchard K, Haskell S, Ferden S, et al. Differences between emergency contraception users in the United States and the United Kingdom. J Am Med Wom Assoc 2002;57:200-3, 14.  Fox J, Weerasinghe D, Marks C, et al. Emergency contraception: who are the users? Int J STD AIDS 2004;15:309–13.  Porter JH. Use of hormonal emergency contraception at a university health centre over a 6 year period. J Fam Plann Reprod Health Care 2001;27:47–8.  Shawe J, Ineichen B, Lawrenson R. Emergency contraception: who are the users? J Fam Plann Reprod Health Care 2001;27:209–12.  Cleland J, Ali MM. Sexual abstinence, contraception, and condom use by young African women: a secondary analysis of survey data. Lancet 2006;368:1788–93.  Heeren GA, Jemmott III JB, Mandeya A, et al. Theory-based predictors of condom use among university students in the United States and South Africa. AIDS Educ Prev 2007;19:1–12.  Raymond E, Trussell J, Polis C. Population impact of increased access to emergency contraceptive pills: a systematic review. Obstet Gynecol 2007;109:181–8.  Shelton JD. Repeat emergency contraception: facing our fears. Contraception 2002;66:15–7.  Raymond E, Halpern V, Levy D. Letters to the editor: “on-demand” oral contraception. Contraception 2009;80:493.  Trussell J, Raymond E. Emergency contraception: a last chance to prevent unintended pregnancy. Princeton: Princeton University; 2010.  Stuart GS, Grimes DA. Social desirability bias in family planning studies: a neglected problem. Contraception 2009;80:108–12.  L'Engle K, Hinson L, Chin-Quee D. I love my ECPs’: challenges to bridging emergency contraceptive users to more effective contraceptive methods in Ghana. J Fam Plann Reprod Health Care 2011;37: 146–51.