What impedes providers from adopting a client orientation? Why do medical barriers persist? How can training be made more effective? In short: What determines family planning provider practices? The answers remain elusive, but in recent years researchers have begun to shine a light in what has been called the "black box"¹ of family planning service delivery.² A variety of tools—in-depth interviews, simulated client studies, and observations—have been used to explain not just "what" providers do, but also "why" they do it. A better understanding of provider behavior is an important key to designing training, supervision and logistic systems that maximize access and quality of care.

Recent attempts at explaining provider behavior have produced illuminating findings, but have only hinted at the constellation of factors that influence service delivery. In this article, we report data from a study in Ghana designed to find out from providers themselves why they impose certain restrictions on clients or engage in medically unnecessary practices.

Context

In 1969, Ghana became one of the first African countries to adopt a population policy. Acceptance of family planning was slow, however, and 20 years later, in 1988, the prevalence of modern contraceptive use had reached only 5%. Between 1988 and 1998, though, the use of modern contraceptive methods nearly tripled, from 5% to 13%. At the same time, the total fertility rate dropped from 6.4 to 4.5 lifetime births per woman.³

To improve quality of care and clients'access to family planning and other reproductive health services, Ghana recently developed and disseminated National Repro- ductive Health Service Protocols (with technical assistance from INTRAH and PRIME).⁴ A few years before, Ghana also had formulated new service delivery policies and standards. All of those guidelines—uniform rules to which all family planning service delivery points should adhere—are designed to remove medical barriers and replace differences between clinical practices with uniform, quality services. Written rules may also protect providers, many of whom fear being blamed by dissatisfied clients or their partners.

Methods

Situation analyses conducted in Ghana in 1993 and 1996 have proven helpful in identifying weaknesses in Ghana's family planning system, by revealing a variety of barriers to access to family planning services.⁵ However, because questions about why certain practices are performed are not included in the standard situation analysis methodology, we conducted a follow-up study in 1994 to learn more about provider rationales for restrictive practices.

To maximize the amount of useful data per site, we used a "negative deviance" strategy, targeting the providers most likely to impose restrictive barriers to services. Using data from the 1993 situation analysis, we created a "barriers to access" index in 1994 for each of the 719 active family planning providers interviewed in that study. The index was simple: A point was added to a provider's score for each reported medical or social restriction to services, including parity requirements, minimum and maximum age limits, laboratory testing requirements, marriage and spousal consent requirements, and others.

Forty-six service delivery points through-out Ghana were chosen as study sites on the basis of high scores from providers at these sites. Three interviewers visited the 46 sites late in 1994 and conducted interviews lasting 1-3 hours. All available providers were interviewed at each site, leading to a final sample of 97 providers.

In each interview, the interviewer first asked the providers a series of simple questions about their clinical practices. A response protocol was then used to probe any responses to the original questions indicating service practices that might be construed as access barriers. For example, providers who said they imposed minimum parity requirements for pill clients were asked their reasons for doing so. Finally, the interviewer reviewed the site's daily register and examined a sample of client records, to validate certain questionnaire responses. For example, the interviewers noted the age and parity range of clients and whether records indicated routine requirements for laboratory tests, pelvic examinations of hormonal method clients or written spousal consent.

This sample was not representative of all family planning providers in Ghana: We deliberately chose facilities where client faced multiple barriers to access. However, we believed that such a sample would be both a cost-effective and rich source of data on commonly reported reasons for enforcing such barriers.

Results

Basic Indicators

The 46 study sites consisted of nine hospitals, 12 health centers, 10 health posts, seven maternity homes, four maternal and child health or family planning clinics, two Planned Parenthood Association of Ghana clinics, a private clinic and a quasi-governmental clinic. They were located in seven of Ghana's 10 regions.

Among the 97 personnel surveyed were 52 auxiliary nurses, 42 professional nurse-midwives, two extension workers and one volunteer. (No doctors were interviewed, as they do not ordinarily provide family planning services at Ghana's public-sector and nongovernmental service delivery points.)

All 46 sites offered combined oral contraceptives, injectables, condoms and spermicides. Forty-four sites offered progestin-only oral contraceptives, 36 IUDs and five (all hospitals) female sterilization services.

Service Restrictions

•Marriage and spousal consent requirements. In clinics visited in both the 1993 and the 1996 situation analyses, about 40% of providers in both surveys reported that marriage was a prerequisite for a client to be offered at least one nonpermanent method.⁶ Spousal consent requirements were even more common.

Providers in our sample of "high barrier" facilities gave a variety of reasons for enforcing marriage requirements. The most common, accounting for one-fourth to one-third of responses (depending on method), was simply that for moral reasons single women should not be allowed to use family planning. Other providers said explicitly that such restrictions discourage indiscriminate sex among unmarried women.

Other rationales for the marriage requirement were specific to various methods. For example, eight IUD providers said they used the marriage restriction to screen out potentially promiscuous IUD clients, presumably to reduce the risk of postinsertion infections. Three providers stated that hormonal methods work continuously and thus are unnecessary for unmarried clients, who may not regularly have sex. One provider worried that unmarried clients might use oral contraceptives to induce abortion.

The most common reasons for spousal consent requirements were variations on the idea that family planning is a decision for both partners or that the husband might oppose family planning or the wife's choice of method (Table 1). Other providers worried about being accused by spouses of providing contraceptives, especially if their use led to side effects, infertility or promiscuity. Some predicted that without spousal consent rules, husbands would accuse wives of infidelity, leading to "punishment," "divorce" and even "fatal consequences." Clients were more likely to face both marriage requirements and spousal consent requirements at government health centers, at health posts, and at maternal and child health and family planning clinics than they were at hospitals and maternity homes (not shown).

•Age restrictions. More than half of providers in both 1993 and 1996 said they enforced minimum and maximum age requirements for various methods. Such restrictions were particularly stringent for injectables. In our purposive sample of service providers, the mean minimum age for prescription of this method was more than 30 years.

When asked their reasons for enforcing minimum age requirements, providers at the "high barrier" facilities most commonly responded with some variation on the notion that the minimum was the age at which women were old enough to marry and have children (Table 1). Some expressed the negative version of this idea—i.e., that family planning should not be allowed for anyone younger than a certain age, or that a minimum age discourages promiscuity, or that women younger than this age are not mature enough for family planning. (Most providers did not differentiate between physical and mental maturity.)

Other reasons given for minimum age limits referred to specific methods and reflected providers' concerns for clients' health. For example, providers seemed to regard injectable contraceptives as particularly risky, and they rationalized age restrictions on this method by citing concerns about infertility or a delay in the return to fertility.

A few providers explained that minimum age restrictions (which, in effect, are also minimum parity restrictions) verify clients' fertility or help them avoid blame from clients who might later prove infertile. Three providers explained that they set age 35 as a minimum for injectable use by saying that "at age 35 clients have children and are nearing menopause, so if they become infertile from injectables, it won't worry them."

Concern for clients' health was also reflected in other rationales. For example, 12 providers claimed that age restrictions were necessary for pills and injectables because young women faced a higher risk of complications from hormonal methods. Two others suggested that when women are younger than 18 or 20, their reproductive organs are too immature for use of hormonal contraceptives.

A few rationales offered were notable for being incorrect or illogical. For example, two providers said that minimum age limits were intended to forestall teenage pregnancies. One stated that progestin-only pills inhibit lactation; another claimed that combined pills, once terminated, increase fertility. Finally, three providers explained their minimum age limit for injectables by saying that the method is dangerous for women younger than 35.

Providers in our sample enforcing maximum age limits for their clients usually cited menopause and health concerns (Table 1). Half of those denying combined pills to older women (usually 35 or older) said that such pills were dangerous or that older clients run higher risks of hypertension or cardiovascular disease. Some also said that "hormonals may lead to early menopause in clients," and that "hidden diseases are exposed after age 35 and may be confused with side effects of the method." In addition, a few thought that some methods (injectables, IUDs, condoms and spermicides) were unacceptable to older clients.

•Parity restrictions. Parity requirements specify that women must have a minimum number of children before receiving a method. These were among the most common eligibility barriers found in the 1993 and 1996 situation analyses; more than half of providers in both surveys claimed to enforce minimum parity rules for at least one nonpermanent method. Interestingly, parity requirements for injectables—sometimes for three, four or even five children—were uniformly more strict than those for IUDs.

Preeminent among the rationales for a minimum parity limit was the belief that hormonal methods, particularly injectables, delay fertility or cause infertility. Among the 80 providers at "high barrier" sites who gave reasons for enforcing parity limits for injectables, nearly all (94%) used some variation of this reason (Table 1). Responses to other questions also indicated that many providers believe injectables cause permanent infertility. One provider said she would not recommend injectables to breastfeeding women because "the client may want to have more children." Another said that "injectables are too strong for the ovaries and destroy them." Dozens of others said or implied that various service restrictions protected the client (or helped the provider escape blame) from infertility caused by injectables.

Delayed fertility or infertility was also the most common reason (three-quarters of stated reasons) for parity limits on the pill (both for progestin-only and for combined formulations). One-third of providers also required that IUD clients have at least one child, and sometimes up to five children. One-third of IUD providers who reported imposing parity restrictions cited possible infertility, while another third mentioned tight cervixes as rationales. Three of the 33 said that the parity requirement helped them escape blame if the IUD left the woman infertile.

The following comments illustrate a few of the seemingly illogical reasons given for IUD parity requirements. One was that "the device causes symptoms of pregnancy such as nausea, so this requirement ensures that such symptoms will not be a new thing to the client." Another provider observed that "the IUD nourishes the uterus, so after removal the client can become pregnant."

Other Service Practices

•Laboratory testing and blood pressure. In the 1993 situation analysis, more than half of providers surveyed indicated that they routinely required laboratory tests of blood or urine before prescribing certain methods. Some researchers have speculated, however, that the prevalence of this barrier may be exaggerated, due to the way the question was posed to providers.⁷ Yet in our visits to the "high barrier" sites (where three-quarters of providers reported routine use of lab tests), review of more than 2,400 client cards revealed that such testing appeared routine in 34 of the 46 locations.

To assess providers' knowledge, interviewers asked respondents who said they routinely required such tests what their purpose was and what were acceptable results or ranges. Most demonstrated a good or adequate knowledge of the rationales for both blood and urine testing and acceptable results or ranges. When asked how they managed cases of abnormal test results, fewer than half said they would provide the method while awaiting confirmation.

As a test of knowledge, interviewers also asked providers why they measured clients' blood pressure, as well as what blood pressure ranges are acceptable for hormonal methods and IUDs. All showed adequate knowledge of the reasons for taking blood pressure. However, more than three-quarters gave upper limits on blood pressure that were lower than those recommended in Ghana's guidelines (160/100), and might therefore restrict access to contraception or unnecessarily reduce a client's choice of methods.

•Follow-up and resupply. Although Ghana's new reproductive health service delivery protocols recommend only one follow-up visit for IUD clients in the first year (in the absence of problems), multiple scheduled revisits for IUD clients may still be a problem in Ghana, as in other countries. Among the 85 IUD providers in our purposive sample, 36% said they required four or more visits during the first year of IUD use, 31% required three and 33% required two or fewer. When asked the reason for multiple IUD revisits, almost all providers explained that they verified that the IUD was in place, checked for complications or counseled the client about infection prevention.

We also assessed oral contraceptive supply and resupply, as Ghana's service protocols specify the number of pill cycles to be provided for new clients (three) and continuing clients (7-13). In the 1996 situation analysis, 84% of all providers said they supplied only one cycle of pills at the initial visit, and more than 80% said they routinely provided three or fewer cycles for revisiting clients who had successfully used the pill for a year or more. In our smaller sample, when asked why they provided fewer than three cycles during follow-up visits, most providers cited the need for frequent health check-ups due to health problems undetectable by clients.

Services for breastfeeding and nonmenstruating clients. When asked which methods they recommended to breastfeeding clients, about 90% of providers at the 46 "high barrier" sites said they recommended progestin-only pills or the IUD. Fewer than half recommended injectables. To assess bias against injectable contraceptives, interviewers asked those not mentioning injectables why they did not do so. About three-quarters said injectables reduce the quantity of breast milk or interfere with lactation.

In the 1996 situation analysis, providers were asked how they managed nonmenstruating clients presenting at the service location for hormonal methods. Half mentioned using pregnancy tests, but even more mentioned sending clients home, with or without a barrier method. Providers in our purposive sample were asked for more details on the various ways in which they ruled out pregnancy for nonmenstruating clients. Most mentioned the use of pregnancy tests. However, only a few noted any of the other means of ruling out pregnancy specified in Ghana's guidelines, including exclusive breastfeeding and client history.

Discussion

Our objective was to learn why providers restrict client access to family planning services in Ghana. Although this was not an in-depth study of the cultural context of providers' personal and professional lives, we obtained some interesting results.

•Why do providers place age, parity, spousal consent, and marriage restrictions on particular methods? Restrictions based on age, parity, marital status and spousal consent are arbitrary, in the sense that most have no medical justification. However, providers in our sample appear to have imposed these restrictions with the best of intentions. Most providers seemed to feel that by doing so, they were protecting both the client and society.

Providers believed they were protecting their clients by denying them access to methods that caused infertility or other health problems. Ghana's new family planning standards and protocols may not address these issues explicitly enough.

A more difficult problem is that providers also wanted to protect their society and culture by preserving traditional values and moral strictures against sex outside of marriage. This debate is common to most cultures, but as health professionals, providers must come to understand that denying women family planning services will not prevent such behavior. Instead, it may expose them to unwanted pregnancies and sexually transmitted diseases.

•Do providers know why they perform certain clinical and laboratory practices and what anomalous results mean? Most had adequate knowledge of the reasons for measuring blood pressure and conducting laboratory tests. A number, though, enforced low upper limits on blood pressure; these might reduce clients' access to effective contraceptives, especially if no distinction is made between estrogen-containing methods and progestin-only methods.

Although blood pressure measurements can be useful in a clinic setting, the utility of requiring laboratory tests is questionable, especially if women who cannot afford such testing are denied services (as was reported by several providers). The most commonly required blood test, that for hemoglobin, is useful when the IUD is being prescribed, but is not indicated for combined pills, which have been shown to increase hemoglobin levels in anemic clients.⁸ Ironically, in the situation analysis, the proportion of active providers requiring hemoglobin tests was greater for hormonal methods than for IUDs. This was unexpected, not only because such tests are less useful for hormonal methods, but also because IUD providers generally work in better-equipped clinics. Other common tests, such as for kidney function (albumin), urine sugar (glucose) and sickle-cell disease, are also of questionable value in this setting.⁹

When lab test results are abnormal, fewer than half of providers in our sample said they would provide a method while the client was treated, counseled or retested. This is unfortunate, since in the absence of other symptoms, the provider's first responsibility to a woman at risk of pregnancy should be to prescribe an effective contraceptive method. Further, it is possible that an abnormal test result stems from a treatable condition, a condition that could be affected more by pregnancy than by contraceptive use, or a false-positive test result. In any case, if retests confirm a diagnosis, or if the client's condition worsens during contraceptive use, the method can always be discontinued without lasting ill effects.

Ghana's new service delivery guidelines use a modified version of the lab test recommendations found in WHO's Medical Eligibility Criteria for Contraceptive Use.¹⁰ Under these new protocols, laboratory testing is not mandatory for any method.

•Why do providers require IUD clients to make so many follow-up visits? Most IUD clients at the sites in our purposive sample were asked to return for revisits three or more times in the first year after IUD insertion. Providers justified multiple revisits by citing the need to verify the presence of the IUD and to check for infection. In recent years, experts have expressed concern that multiple IUD revisits are not useful or cost-effective, that they exacerbate overcrowding at some clinics and that they may discourage potential IUD acceptors.¹¹ Ghana's current recommendation, therefore, is that safe IUD provision need only include a one-month follow-up visit and careful counseling of clients to return if problems occur.

•Why are oral contraceptive clients not given more pill packets per visit? Many providers seem convinced that oral contraceptive clients need frequent check-ups to avoid serious medical problems. For this reason, eight in 10 providers in the most recent situation analysis gave only one pill packet at the first visit and three or fewer at subsequent visits.¹² Yet side effects from the pill are almost always minor, and almost any problem, minor or major, can be detected by a well-counseled client for whom the pill was indicated at the initial examination.

•Why is contraception not provided to nonmenstruating new clients? Denial of services to clients who are not menstruating is a serious barrier to family planning services around the world.¹³ We found that when providers were asked how they ruled out pregnancy among nonmenstruating clients, most mentioned only pregnancy tests. This indicates that providers did not know about, or were uncomfortable with, other means of ruling out pregnancy, such as a client history.

Conclusions

Providers in our nonrepresentative sample showed inadequate technical knowledge of the precautions against and side effects of modern contraceptive methods. Their goal of protecting their clients is admirable, but in exaggerating the dangers of contraceptive use, they may be doing more harm than good. Injectables are especially misunderstood. As has been noted in other African countries,¹⁴ many providers seem to have exaggerated notions of injectable-related infertility. Further, age and parity requirements for injectables are particularly burdensome, and few providers recommend injectables to breastfeeding clients.

Moreover, in imposing personal values and mistrusting their clients, providers demonstrate a want of empathy and professionalism. The "end" of improving society's morals does not justify the "means" of denying services to unmarried clients and of speculating on the fidelity of married ones. Similarly, it is unfortunate that providers do not seem willing to trust nonmenstruating clients, for whom ruling out pregnancy often requires a measure of trust.

Finally, providers in this sample seem to rely too much on laboratory testing. We hope that with the advent of new guidelines, the time and money spent by clinics and clients on such testing is now better used elsewhere.

Our findings are based on a sample of Ghanaian family planning providers that was not random; providers who impose many restrictions on clients were specifically targeted. We cannot therefore extrapolate the prevalence of barriers in our sample to Ghana as a whole. However, situation analyses of family planning continue to show that these barriers are a problem in Ghana (overall, findings were worse in 1996 than in 1993¹⁵), and the reasons that providers used to explain their practices are common ones.

If these barriers to client access indeed are widespread, it is commendable that Ghana's new protocols and standards for reproductive health service delivery address many, if not all, of these restrictive practices. We hope that these newly introduced guidelines and the training that has accompanied their dissemination will go far in reducing the unnecessary hurdles faced by family planning clients in Ghana.