METHODS

From a list of all (approximately 1000) registered pharmacies provided by the Mexico City Chamber of Commerce, we randomly selected 5% (n = 57) using a pseudorandom number generator (SPSS, Version 7.0, Chicago, IL, USA). Eleven of the selected pharmacies refused to participate (response rate of 81%). We randomly chose 11 additional pharmacies to replace them. After a small pilot (n = 5), we made minor changes to the questionnaire, mainly to reduce its length. Because the pilot survey did not otherwise differ from the main study questionnaire, we analysed all data together. All interviews took place in August 1999.

We designed the survey to simulate the experience of individuals with STI symptoms seeking assistance from pharmacies. Working in Spanish, an investigator interviewed the first available attendant at each pharmacy. Each individual provided oral informed consent before participating. We notified attendants that we were investigating the diagnosis and treatment of sexually transmitted vaginal and genital infections, and we offered 20 pesos (~US$2) as compensation. Approximately 50% of individuals accepted this compensation.

The survey (available upon request) consisted of three parts. The first section collected information about the pharmacy’s clientele, including typical monthly client number and breakdown by sex, type of health seeking behaviours (self medication versus requests for pharmacist diagnosis and treatment recommendation) and typical client descriptions of sexually transmitted vaginal or genital infection symptoms.

The second section investigated pharmacy attendants’ diagnosis and/or treatment recommendations, based on hypothetical case studies, for three sets of infection symptoms (table 1). Case study 1 described symptoms suggesting gonorrhoeal and/or chlamydial infection; case study 2 depicted herpes simplex virus (HSV) infection; and case study 3 described symptoms of trichomoniasis and/or bacterial vaginosis (BV) infection. We organised infection information on flashcards and presented them in random order for each case study. We then asked participants to explain how they might proceed with hypothetical clients in each case. Attendants occasionally sought advice from colleagues, which was permitted under the protocol.

The final section collected background information about the pharmacy and the participating attendant.

Analysis

We analysed participants’ responses to the case studies to determine which individuals properly diagnosed and/or treated the specific infections we described, or, those who properly diagnosed and/or treated the STI syndrome. We considered participants to have answered correctly if they suggested at least one appropriate diagnosis or treatment. We did not penalise participants who gave multiple suggestions, even if some of the treatment recommendations were ineffective or irrelevant.

Our analysis scheme reflects current STI diagnosis and treatment recommendations, as well as the debate between conflicting models. Syndromic STI management, which has proved to be valid, feasible, and cost effective for certain syndromes, is based on the identification of consistent groups of symptoms and easily recognised signs (syndromes), and the provision of treatment that targets the majority or most serious organisms responsible for producing that syndrome.²⁰ Syndromic treatment has many advantages. Since an individual who is positive for one STI is often also harbouring another infection, treating his/her symptoms as though they are caused by numerous potential pathogens may be a more efficient therapeutic option than treating just the “best guess” pathogen. Clinicians in many parts of the world (including the pharmacy attendants surveyed here) may not have access to microbiological laboratory assays that confirm the presence or absence of a specific pathogen. Syndromic treatment is also useful when certain drugs are effective against more than one infection (for example, antibiotic azithromycin, which effectively treats both gonorrhoea and chlamydia).

Syndromic STI treatment, however, also has several drawbacks. One study assessing the effectiveness of syndromic STI management found substantial overtreatment,²¹ and another project reported that self and health worker recognised STI symptoms exhibited low sensitivities, specificities, and positive predictive values in STI diagnoses.²² Though the World Health Organization (WHO) recommends syndromic management, it also cautions that certain treatment algorithms are limited, particularly those reporting on the management of vaginal discharge and cervical infections in women.²⁰

We checked participants’ treatment recommendations against those recommended by WHO,²⁰ a leading reproductive health textbook, Contraceptive Technology,²³ the 1998 edition of the Physicians’ Desk Reference²⁴ and its 2002 Mexican equivalent, PLM-Diccionario de Especialidades Farmacéutica,²⁵ as well as Mexican doctors who treat STI patients through their medical practices.

RESULTS

Pharmacy personnel characteristics

Pharmacy attendants varied widely in age and experience. Mean age was 32 years (range 18–60) and median time working at that pharmacy was 5 years (range 0.3–48.0 years). Half the participants were men, half women; 11% had some type of pharmacy certification. Fully licensed pharmacists were ever available in only 54% of pharmacies. The interviews lasted approximately 12 minutes.

Typical pharmacy clients’ complaints

Attendants saw an average of 28 clients per month with symptoms of vaginal or genital infections. Nearly two fifths (39%) of participants reported seeing fewer than 10 clients per month, while half (51%), saw between 10 and 50. The remaining 11% reported seeing over 50 clients per month (percentages do not add up to 100% because of rounding). Attendants estimated that over 80% of STI consultations were with women.

Although more than half the attendants (58%) reported that some clients were embarrassed to describe symptoms of vaginal or genital infections, only two pharmacies had designated places where clients and attendants could speak privately.

According to attendant reports, 64% of presumptive STI clients already knew what medication they wanted to buy; these self directed clients were usually filling doctors’ prescriptions or incorporating information from television commercials or other media. The remaining 36% of clients typically requested treatment recommendations based on their symptoms.

We asked survey respondents to describe clients’ “typical” symptoms. Participants most frequently cited itching, vaginal discharge, and burning sensations as complaints from women, with only one participant not suggesting any female symptom. Nearly half the respondents (47%), however, could not describe any “typical” STI symptoms reported by male clients. Of those who answered, itching, irritation, and penile lumps or bumps were the most commonly reported complaints.

Hypothetical case study results

We presented flashcards displaying three sets of typical STI symptoms (table 1). Case study 1, describing symptoms of gonorrhoea and/or chlamydia, and case study 2, describing HSV, referred to hypothetical male clients. Case study 3, depicting symptoms of trichomoniasis and/or BV, described a female client. Table 2 summarises participants’ responses.

Sixteen participants (28%) gave one or more diagnoses (not necessarily correct) for the gonorrhoea/chlamydia case, 12 gave diagnoses in the HSV case (23%), and 23 attendants (40%) attempted to diagnose the trichomoniasis/BV case. Accurate diagnoses, however, were rare. Participants suggested gonorrhoea or chlamydia in less than one third of attempts (n = 5; all suggested gonorrhoea). Attendants correctly diagnosed HSV in slightly more than half of their attempts (n = 7). No participants recognised the trichomoniasis/BV syndrome beyond the general diagnosis of “vaginal infection.”

Table 2 also describes recommended treatments. Participants suggested treatment in the gonorrhoea/chlamydia case on 24 occasions (42% of the time), and nine antibiotic prescriptions were appropriate for gonorrhoea: spectinomycin (recommended four times), ceftriaxone (twice), rosoxacin (once), ciprofloxacin (once), and pefloxacin (once). No participant suggested azithromycin (effective against both gonorrhoea and chlamydia), WHO’s suggested treatment based on the described syndrome. In 11 of the incorrect prescriptions, participants suggested antibacterials (such as chloramphenicol) intended for treatment of urinary tract infections (UTIs) or antimicrobials (such as phenazopyridine) for treatment of pain, burning, and discomfort caused by infection or irritation of the urinary tract (though not effective against the infection itself). The remaining inappropriate recommendations included drugs intended for other reproductive tract infections, including the antifungal clotrimazole (effective against candidiasis) and the antiprotozoal tinidazole (effective against trichomoniasis).

Participants prescribed medication for the HSV case study 19 times (33% of all attendants). Attendants suggested appropriate antiviral treatments on seven occasions: aciclovir (five times), ribavirin (once), and tromantadine (once). (Ribavirin is a broad spectrum, ribonucleuic acid (RNA) virus mutagen. Though used most commonly for treatment of respiratory syncytial virus (RSV), ribavirin is also effective against HSV, hepatitis C virus, measles, mumps, Lassa fever, and other viral pathogens.) Incorrect treatment recommendations covered a broad range, including medications more appropriate for gonorrhoea (antibiotics spectinomycin, rosoxacin, and ceftriaxone), either gonorrhoea or chlamydia (antibiotic azithromycin), UTIs (antibiotic ampicillin), and candidiasis (antifungal clotrimazole).

Attendants were much more confident when suggesting treatment for the trichomoniasis/BV case, as 83% of all respondents made treatment recommendations (n = 47). However, participants suggested appropriate treatments only nine times (16% of all attendants). All nine mentioned antiparasitics metronidazole or secnidazole, effective against both trichomoniasis and BV. On 34 occasions, participants suggested treatment effective against candidiasis (antifungals clotrimazole or miconazole, or both together). The remaining ineffective products were more appropriate for urinary tract infections (UTIs) (antibiotic ampicillin) or UTI symptoms (antimicrobial phenazopyridine).

Participants suggested physician follow up 38 times (67% of all participants) following the gonorrhoea/chlamydia case study. Most (n = 30) had not recommended treatment, though on three occasions, neither treatment nor physician follow up was recommended. Similarly, participants recommended physician follow up on 37 occasions (65% of all participants) for the HSV case study, and all but one of these had not suggested treatment. Three attendants (the same three participants as in the gonorrhoea/chlamydia case) recommended neither treatment nor physician follow up. Attendants recommended physician follow up for the trichomoniasis/BV case 15 times (26%), with 13 of these suggesting follow up in the first two case studies as well. On two occasions, participants recommended neither treatment nor physician follow up.

DISCUSSION

Our results suggest that Mexico City pharmacy attendants possess inadequate knowledge to provide accurate STI diagnoses and treatment. These self reported data indicate poor training and high failure rates in identifying and recommending appropriate treatments for three sets of common STI symptoms.

Aside from inadequate training, the potential reasons for low recognition and mistreatment of STIs are many. One interpretation is that even when participants did not have useful treatment recommendations, they suggested costly but ineffectual (for STI treatment) medications in order to make a sale. Alternatively, others may have recognised general symptoms of reproductive tract infections and recommended what came to mind first, possibly influenced by advertising campaigns. Some respondents may have simply listed possible treatments because they did not know what else to recommend.

As might be expected, diagnoses did not always precede treatment recommendations, nor did treatment recommendations always follow diagnoses. That is, some attendants made diagnoses without prescribing medication, and others prescribed medications when they were unwilling to specify a diagnosis.

Some participants (12%) reported that they only attended individuals with doctors’ prescriptions. These pharmacies may not therefore constitute a source of primary health care for their clients (beyond filling clients’ prescriptions). For the remainder of pharmacies, it is possible that some clients have prescriptions some of the time. However, we estimate that perhaps 50% of individuals are seeking medical treatment for vaginal or genital infections from pharmacies without a medical doctor’s advice. This supports the major role of pharmacies in the provision of primary health care for STIs.

Our study has several limitations. Our results attempt to characterise the competence of pharmacy practice based on only three STI case scenarios and only within Mexico City. Since we interviewed the first available pharmacy attendant, we approximated the experience of individuals who do not specifically ask to consult trained pharmacists, so our survey cannot appraise the performance of certified pharmacists. This study also did not address the conflicting roles of the pharmacy as merchant and medical provider, nor the impact of commercial drug company publicity on employees. Finally, our results reveal that many more women than men approach pharmacies for STI diagnosis and treatment; unfortunately, two of the three hypothetical case studies referred to men. Attendants’ greater experience “treating” women may have resulted in more successful responses to our case studies if more of the descriptions had referred to women.

These data suggest an obvious policy recommendation: implementation of aggressive and sustained training for pharmacy personnel. Other studies have shown that improvements in appropriate diagnosing and treatment are possible following such training efforts, but these same reports caution that the improvements are difficult to maintain.^9,26,27 In addition, intensified training entails thorny legal and logistical decisions. For example, programmes that educate pharmacy personnel about appropriate syndromic STI management must carefully address prescribing regulations, since such “prescribing” is technically illegal. This type of training programme may also meet with opposition from physicians, who want to minimise competition from outside medical alternatives. In contrast, training efforts that encourage pharmacy attendants to refer all potential STI clients to a physician are also limited, since individuals who seek pharmacy treatment are often those who cannot afford physicians and since the efforts would ask pharmacists to act against their financial incentives. Such a programme might have the undesired consequence of reducing the total number of people who obtain appropriate STI treatment.

Any training effort must carefully balance these challenges, though clearly, dissemination of accurate STI information to the Mexican population as a whole, as well as a broad provision of educational materials to pharmacists specifically, is essential. Such information can enable people to avoid initial infection, as well as allow both individuals and pharmacy attendants to assess better the severity of disease condition and possible courses of action. These strategies could help prevent misdiagnoses, mistreatments, and potential sequelae due to pharmacies’ ineffective STI treatment.