Quality of Sexually Transmitted Disease Case Management in Nairobi, Kenya

A Comparison Among Different Types of Healthcare Facilities

Abstract

Background 

In Nairobi, the prevalence for sexually transmitted diseases (STDs) among attenders at antenatal and family planning clinics is substantial, but knowledge about the quality of STD case management is scarce.

Goal 

To assess quality of STD case management in Nairobi healthcare facilities.

Study Design 

All the facilities in five sublocations were enumerated. In 142 facilities, 165 providers were interviewed, observed during 441 interactions with patients who had STDs, and visited by a simulated patient.

Results 

For observations of patients with STDs, correct history-taking ranged from 60% to 92% among the various types of facilities, correct examination from 31% to 66%, and correct treatment from 30% to 75%. The percentage of correctness for all three aspects (World Health Organization prevention indicator 6) varied between 14% and 48%. Public clinics equipped for STD care performed best in all aspects, whereas treatment was poorest in pharmacies and private clinics. The providers trained in STD management performed better than those without training.

Conclusions 

Quality of STD case management was unsatisfactory except in public STD-equipped clinics. This indicates the need for improvement by interventions such as further training in syndromic management, improved supervision, and the introduction of prepackaged syndromic management kits.

SEXUALLY TRANSMITTED DISEASES (STDs) are a major burden of disease in many developing countries. ^1,2 Additionally, they enhance the transmission of HIV. ^3,4 Treatment of STDs is therefore both a goal in its own right and a strategy for the prevention of HIV infection. ⁵ Unfortunately, the quality of STD case management in developing countries often is unsatisfactory. ^6–8

In Kenya, the prevalence for STDs in the general Nairobi population is unknown. Studies among subpopulations have shown STD infection rates to be substantial. ^9–11 Although STD prevalences among female attenders at antenatal and family planning clinics in Nairobi declined between 1992 and 1999, they still were as high as 2% for gonorrhea, 3% for syphilis, and 17% for chlamydia in 1999. During the same period, HIV prevalence increased from 15% to 19%. ¹² Knowledge about the quality of STD case management in Kenya is scarce. One study was conducted in a rural area (Kisumu district, unpublished), but not much is known about the capital, Nairobi, where patients with STDs seek care in the public, private, and informal sectors. ¹³

This report describes the results of a study to assess the quality of STD case management in the healthcare facilities of five Nairobi sublocations. Provider performance was assessed through patient observations, interviews, and simulated patient visits. Six types of facilities are distinguished and compared. In the private sector, pharmacies, mission clinics, and nongovernmental organization (NGO) or community-based clinics (operated by international NGOs such as the Crescent Medical Aid or by community self-help groups) are distinguished from private clinics owned by individuals (hereafter termed “private clinics”). In the public sector, a distinction is made between clinics equipped and those unequipped for STD care (locally termed strengthened and nonstrengthed clinics, respectively). Equipped or strengthened clinics, which serve as STD referral clinics for other public facilities, are clinics staffed by providers trained in syndromic management by the Ministry of Health (supported by donors such as the Belgian government and CIDA) in the mid-1990s, in which a regular supply of STD drug kits was established. Besides comparisons between different types of facilities, providers trained in STD management were compared with untrained providers, and doctors were compared with nurses and clinical officers.

We know of no previous studies comparing different types of facilities by the quality of their STD management. Most studies in developing countries have focused exclusively on clinics in the public sector, ^7,8,14,15 on pharmacies, ^16,17 or on the private sector. ^18,19 Two studies included both public and private clinics, but did not compare their performance. ^6,20 Only a few studies have used simulated patients in assessing quality of STD care. ^7,16,17,21

In this report, quality is defined as the degree to which STD case management is correct according to World Health Organization (WHO) criteria during provider–patient observations. These WHO criteria are described in a protocol for the assessment of STD case management, which has been used in most studies on quality of STD care. ^{6–8,14,15,20–22} The aspects of STD case management considered in this report are history-taking, physical examination, and treatment. Education aspects such as condom promotion, contact-tracing, compliance, and counseling (the four C’s) are reported separately. ²³

Methods

Enumeration

Five densely populated sublocations in Nairobi were selected for this study: Mathare and Korogocho in the north, Buruburu in the east, Kibera in the south, and Kawagware in the west (Figure 1). These areas contain approximately one third of the inhabitants in Nairobi, a city with an estimated population of 2.5 million. The sublocations were selected to include a mix of medium, low, and very low socioeconomic areas, and to be spread geographically over the city. ²⁴ Within each sublocation, the researchers performed an enumeration by walking through all the streets and listing the facilities encountered. In each facility a short questionnaire was administered to assess the type of facility, the number of patients with STDs seen in the previous week, and some other background characteristics of the facility. The inclusion criterion required that four or more patients with STDs had been seen in the preceding week. When this criterion had been met, consent was asked from the head of the facility for its participation in the study.

Fig. 1: Schematic map of Nairobi, with the location of the five sampled areas, the city center, and the public clinics equipped for sexually transmitted diseases. The overview shows the number of studied facilities in each area per type of facility.

Facilities in neighboring areas and the city center that inhabitants from the selected sublocations visited frequently for STDs also were studied if they met the inclusion criterion. These facilities were identified in a community study on STD-related healthcare-seeking behavior, which was part of the same research project. Additionally, of the 45 Nairobi City Council public clinics, the 10 equipped for STD care were included in the study, along with three university research clinics and the Nairobi STD referral clinic (Figure 1). Within each facility, all the providers present on a randomly chosen day were studied.

Provider–Patient Observations

Providers were approached by a research assistant (a qualified nurse or senior medical student), who observed their patient interactions throughout one full working day. In cases of STD, the consent of the patients was asked for the research assistant to be present during the consultation and physical examination. For each patient with STD, research assistants recorded all the actions of the provider and the background characteristics of the patient using a standard questionnaire based on the WHO protocol. ²²

Interview

Interviews were used to assess providers’ knowledge of STD case management. ²² These interviews followed a structured questionnaire adapted from the WHO protocol. They covered background characteristics of the provider (profession, training), the usual steps taken in the management of STDs, and the first-choice treatment given for particular syndromic and etiologic diagnoses. The interview was conducted after the observations so it would not influence the performance of the providers during the observations.

Simulated Patient Visit

The simulated patient visit was used to assess the actual daily practice of providers, as opposed to “optimal performance,” which was measured during patient observations when providers knew they were being studied. Male and female research assistants were trained to act as simulated or dummy patients with standardized STD conditions (white milky discharge, no pain) and a recent history of unsafe sex. The simulated patients visited the providers at an unannounced time and bought the drugs the providers prescribed. The providers were not aware that the patient was not a real patient. Because the research assistants did not have actual STD symptoms, they avoided or refused being examined, allowing any specimen to be taken for tests, or taking drugs on the spot by injection or otherwise.

Directly after the visit, the simulated patients completed a standardized questionnaire, recording all the steps undertaken by the provider during the diagnostic process and the drugs prescribed or sold. Where possible, the simulated patient visit preceded the observations and interview so these would not influence the performance of the provider during the simulated patient visit.

Data Collection and Analysis

Data collection took place from February until May 1999. Problems were encountered in finding the same provider at different times. As a result, not all the providers were studied with all three methods. Difficulties inherent to simulated patient method were anticipated and addressed during the training of the research assistants (i.e., how to avoid or refuse being physically examined, allowing samples to be taken, or being forced to take drugs at the spot). The method proved to be feasible, with only one provider expressing doubts about the genuineness of the simulated patient.

All interviews, observations, and simulated patient forms were entered daily into Epi Info 6. As an exception to WHO protocol, female patients with reports of discharge were included in the analyses of provider–patient observations. Analyses consisting of cross-tabulations, χ² calculations, and logistic regression analyses for calculation of 95% CI were performed with SPSS PC Version 8.0. Some providers saw more than one patient during the observations, so these observations were not statistically independent. Therefore, χ²tests during the observations were adjusted to account for correlations between observations that result from the clustering of patients with providers. To accomplish this, the Generalized Estimated Equations (GEE) approach in logistic regression of Proc Genmod in SAS Version 6.12 TS level 0060 was used.

Outcome Measures for Observations of Patients With STD

Correct history-taking.

According to WHO, correct history-taking involves asking questions on the nature of symptoms, the onset or duration of symptoms, and recent sexual contacts. ²² Besides these aspects, we also assessed whether a provider asked a patient about the history of previous STDs and about the history of care-seeking for the current problem.

Correct examination.

Correct examination of patients with STD involves fully exposed genitals, with females lying down. Patients should be examined subsequently for discharge and lesions, with the foreskin of uncircumcised male patients retracted and the labia of female patients separated and inspected. ²² Additionally, we assessed whether a provider performed a speculum and/or a bimanual examination on the female patients. Because pharmacies have only a drug-dispensing role, they were excluded from the analysis regarding correctness of examination.

Correct treatment.

In cases for which a provider relied on laboratory test results for diagnosis, treatment was assessed using guidelines for etiologic diagnosis (i.e., the determined causative agent had to be treated correctly). In all other cases, correct treatment was assessed using guidelines for syndromic diagnosis. Syndromic management of STDs was introduced in Kenya during the mid-1990s by training providers mostly from the public sector, and by developing and widely distributing a syndromic management flowchart. According to the flowchart, for vaginal discharge in female patients, trichomonas and candidiasis must be treated first. If symptoms do not disappear, gonorrhea and chlamydia must be treated. In our assessment, we considered both therapies correct, regardless whether the patient was a first-time or follow-up patient. For urethritis in male patients, gonorrhea and chlamydia had to be treated correctly. Correct treatment of syphilis and chancroid was required for genital ulcer disease in male and female patients. In cases of incorrect syndromic treatment, we also recorded whether one causative agent was treated correctly (i.e., clinical management without laboratory tests).

Correct treatment was assessed using national and WHO guidelines as criteria. ^25,26 Treatment that adhered to either of these guidelines was considered correct, with abundant doses and durations also regarded as correct. Short doses and durations were considered incorrect unless they proved sufficient to clear the etiology or syndrome according to East African guidelines. ²⁷ Drug regimens with developed resistance in Kenya, as determined by national/East African standards, were considered incorrect, even if they fulfilled WHO guidelines. ²⁷ Providers were assessed for the drug therapy they prescribed, even if the patient did not have enough money and left the clinic with fewer drugs than prescribed.

Prevention indicator 6.

In the WHO protocol, prevention indicator 6 (PI6) is described as the proportion of patients with STDs for whom history-taking, examination, and treatment are all correct. ²² Because pharmacies are not assessed for correct examination, PI6 could not be calculated for this type of facility. It should be noted that for PI6 to be correct, examination had to be performed, regardless whether a provider had given etiologic or syndromic treatment.

Results

During the enumeration, 201 healthcare facilities were identified. Of these facilities, 23 refused to be studied before their eligibility was assessed, 29 were not eligible (i.e., were visited by fewer than four patients with STDs during the week before the enumeration), and 7 did not give consent for the study. Altogether, 142 of 168 eligible facilities were studied (Figure 1), representing a response rate of 85% if a similar rate of eligibility is assumed for the facilities that refused. Refusal was highest among pharmacies.

A total of 192 providers were included in the study, averaging 1.4 providers (range 1–4 providers) per facility. Because not all providers could be studied with each method, 138 were studied with all three methods, 27 only with interview and observation, and 27 only with a simulated patient visit. Of the 165 providers observed during patient interactions, 27 did not see a single patient with STD during a full working day. In all, 441 patients with STDs were observed, averaging 2.7 patients (range 0–19 patients) per provider.

The distribution of facilities, providers, and patient observations over the different types of facilities is given in Table 1. Private clinics owned by individuals account for approximately half of the total facilities and providers, but only one fourth of the total patient observations. Public clinics equipped for STDs and mission clinics see many patients per provider. They each provide approximately one tenth of all the facilities and providers, but one fourth of the patient observations. Public clinics not equipped for STD care comprise the smallest sample size in terms of facility, provider, and patient observation. Because they normally refer patients with STDs to the equipped clinics, they did not meet the study inclusion criterion of having managed at least four patients with STD in 1 week.

Table 1: Number of Facilities, Providers, and Patient Observations by Type of Facility

Of the observed providers, 49% were nurses, 18% clinical officers, 15% doctors, and 11% pharmacists/pharmacy technologists. One third of the providers had taken an in-service course on STD management. Of the observed patients with STDs, most were married (61%) females (59%), 20 to 29 years of age (58%), who had problems involving discharge without ulcers (61%). ²⁴

Comparison of Facilities During Observations of Patients With STDs

Table 2 shows the proportion of the 441 patients with STDs who were managed correctly in terms of history-taking, examination, and treatment. The score ranges for the different type of facilities also are shown. The first part of Table 3 gives the WHO summary scores for history-taking, examination, treatment, and PI6 for the different types of facilities.

Table 2: Quality of Sexually Transmitted Disease (STD) Case Management During Patient Observations (n = 441): Items and Summary Scores for History-Taking, Examination, and Treatment^*

Table 3: Quality of Sexually Transmitted Disease (STD) Case Management During STD Patient Observations, Interviews, and Simulated Patient Visits: Summary Scores for History-Taking, Examination, Treatment, and PI6, by Type of Facility^*

Correct history-taking.

Fairly high scores were obtained for the three variables in the WHO criteria for correct history-taking: inquiry about the nature of symptoms, their onset or duration, and recent sexual contacts (Table 2). The performance on inquiry about previous STDs and care-seeking for the current problem was worse, with no type of facility scoring higher than 50% for the latter variable. The WHO summary score for correct history-taking was between 60% and 92%, with no significant differences between types of facilities. There were no significant differences between the management of male and female patients, between the performance of male and female providers, or between the performance of providers on same or opposite sex patients for any of the variables involved in correct history-taking.

Correct examination.

For physical examination, the scores generally were lower than for history-taking. Approximately two thirds of all the patients with STDs were physically examined, and about half were examined correctly according to WHO criteria (Table 2). Differences between types of facilities were not significant. Overall, male patients were examined physically more often than female patients (76% versus 65%;P = 0.015), regardless of the provider’s gender. For female patients, a speculum and bimanual examination were performed rarely (10% and 24% of cases, respectively), especially in mission clinics (0% and 2%, respectively).

Correct treatment.

Approximately 30% of the patients with STDs were treated on the basis of an etiologic diagnosis (Table 2). This proportion varied largely between types of facilities, with mission and private clinics having the highest proportions (60% and 35%, respectively), and pharmacies the lowest (3%). Overall, 54% of the patients were treated correctly, with a large variation between facility types (P < 0.001): both types of public clinics performed best (75% and 71%, respectively), whereas pharmacies and private clinics performed worst (38% and 30% respectively, first part of Table 3). Although syndromic treatment contributed more to the proportion of correct treatments than etiologic treatment (33% versus 21%, respectively;Table 2), the providers performed better when treatment was given on the basis of an etiologic diagnosis than when it was based on a syndromic diagnosis (68% versus 47%, respectively). Regarding incorrect syndromic management with one agent treated correctly, the proportion was highest for pharmacies and private clinics (55% and 41%, respectively), and lowest for public clinics equipped for STD care (7%).

Prevention indicator 6.

The overall score for PI6 was 27%, meaning that only 27% (95% CI, 23–32%) of the 441 observed patients with STDs were managed in an appropriate way. There was a large variation between facility types (P = 0.01). Public clinics equipped for STD care had the highest PI6 score (48%), followed by NGO or community-based clinics (28%). The remaining three types had scores between 14% and 18%. These low scores were mostly the result of incorrect examination or incorrect treatment.

The facilities on which the results of Table 2 and 3 were based did not precisely represent the sampled study areas because STD-equipped public clinics from all over Nairobi were included, as well as facilities in the city center or those neighboring the study areas (see Methods section). To check the robustness of our results, we tried to derive more precise estimates representing STD care for people from the study areas. Therefore, we weighted city center/neighboring facilities with a factor 0.333 because the population size of the study areas was approximately one third that of Nairobi. We also excluded the STD-equipped public clinics not visited frequently by people from the study areas. ²⁴ As a result, the summary score for history-taking remained at 81%, whereas the score shifted from 51% to 46% for examination, from 54% to 47% for treatment, and from 27% to 21% for PI6. Therefore, the total scores given in Table 2 and the first part of Table 3 are slightly optimistic for the study areas. This is the case also for Nairobi as a whole, because the study areas can be seen as a representative sample of Nairobi.

Comparison of Facilities During Interviews and Simulated Patient Visits

The middle and last parts of Table 3 show the results of the interview and simulated patient methods. Some outcome measures were calculated a bit differently than for the observation method. For the interviews, providers were asked to give the first-choice treatment they usually prescribed for four different STD etiologies (gonorrhea, chlamydia, primary syphilis, and chancroid) and four STD syndromes in absence of a definitive etiologic diagnosis (urethral discharge, vaginal discharge, and genital ulcer in male and female patients). Correct treatment was assessed for each of these eight conditions and summarized by the proportion correct for each provider. For the providers who said they normally give only etiologic or only syndromic treatment, only the four relevant conditions were summarized. The final treatment score thus reflects the average knowledge of a provider for the approach he or she normally uses. Subsequently, PI6 was calculated using the proportion score for treatment. For the simulated patients, the only possible assessment was whether a provider requested an examination or not because these patients had to avoid physical examination. This information was used to calculate PI6.

During the interviews, there were significant differences between types of facilities for correct treatment and PI6. The STD-equipped public clinics again scored highest, whereas the pharmacies and private clinics again scored lowest (Table 3 [middle]). The same was true for correct treatment during the simulated patient visits (Table 3 [bottom]). There also were significant differences during the simulated patient visits for history-taking, with the pharmacies scoring the lowest, followed by the mission clinics.

When the three methods for each type of facility were compared, a pattern could be seen for history-taking, examination, and PI6: The interview scores were higher than the observation scores, whereas the simulated patient scores generally were the lowest. Therefore, the knowledge of providers during the interviews was adequate, whereas performance during the observations was less optimal, and daily practice during the simulated patient visits was the least. For history-taking, the discrepancy was largest for the pharmacies (76%, 60%, and 31%, respectively), whereas for the examination, the discrepancy was largest for the private clinics (89%, 52%, and 33%, respectively) and NGO or community-based clinics (90%, 49%, and 29%, respectively).

Concerning treatment, the differences between the three methods were relatively small. Both types of public clinics, which showed good knowledge during the interviews, also performed well during daily practice. The pharmacies and private clinics, with little knowledge during the interviews, also performed rather poorly during daily care.

Comparison of Providers During All Three Methods

Besides comparing types of facilities, it was interesting to compare providers. Table 4 compares the providers who followed an in-service course on STD management for all three methods with those who did not. An in-service course on STD management significantly improved knowledge of the providers on history-taking, treatment and PI6 (interviews), performance on examination, treatment and PI6 (observations), and daily practice during treatment (simulated patient visits, Table 4).

Table 4: Quality of Sexually Transmitted Disease (STD) Case Management During STD Patient Observations, Interviews, and Simulated Patient Visits: Comparison of Providers Who Attended an In-service Course on STD Management With Those Who Did Not Regarding Summary Scores for History-Taking, Examination, Treatment and PI6^*

Providers also can be compared on the basis of their profession. When doctors (n = 69) were compared with nurses (n = 228) and clinical officers (n = 86) during the observations, the doctors performed significantly better in terms of correct examinations (71% versus 45% and 53%;P = 0.02). This caused them to score slightly, but not significantly, higher on PI6 (37% versus 24% and 30%;P = 0.10). Regarding treatment, which is the most clinical aspect of STD case management, doctors did not perform any better than nurses or clinical officers (55% versus 57% and 56%;P = 0.25). During the interviews and simulated patient visits, no significant differences among doctors, nurses, and clinical officers were observed.

Because the providers were not aware that they were being studied by simulated patients, this method was appropriate for assessing whether patients with STDs are stigmatized or made to feel unwelcome (also called “the rejection phenomenon”). In the 165 simulated patient visits in our study, only 5% of the patients said that the provider was unfriendly or seemed to despise or condemn them for contracting an STD. The proportion was highest for the public sector (13% for STD-equipped and 25% for STD-unequipped clinics), but the numbers were very small for drawing conclusions (2 of 16 cases and 1 of 4 cases, respectively). Privacy seemed to be a bigger problem, with 19% of the simulated patients indicating a lack of privacy. When pharmacies (no privacy indicated for 38%) were excluded, 15% of the simulated patients still indicated a lack of privacy.

Discussion

In this report, three methods for assessing the quality of STD care are compared: observation of provider–patient interactions, interviews, and simulated patient visits. Several aspects must be taken into account when the results of the three methods are compared.

First, not all the providers who were observed and interviewed were also visited by a simulated patient, and vice versa. However, the findings were the same when the analysis was restricted to the 114 providers studied with all three methods. Second, although there was only one score per provider for the interview and simulated patient method, the number of patient observations varied between providers. This means that the providers with many patients had a larger share in the observations than in the interviews and simulations. However, the results are similar when an average observation score is used for each provider. Third, the simulated patients were not quite comparable with the observed patients because they presented with a standard history of STD problems and unsafe sex, without any vagueness, complications, or inconsistencies. Furthermore, they refused examination, sample-taking, and injections. Therefore, they might have been treated differently than the average observed patient. This means that the observed differences in performance between the simulated and real patients might have resulted from the patients not being comparable. However, differences between real and simulated patients would not be expected to play a role in history-taking. The fact that even history-taking was scored better in real patients than simulated patients (81% versus 62%) indicates that providers score better during observations because they know they are being observed.

According to the observations of patients with STDs in this study, the overall quality of STD case management in Nairobi is unsatisfactory, except in public clinics equipped for STD care. The study findings showed physical examination and treatment to be most problematic, causing the overall WHO indicator for STD management PI6 to be lower than 50% for all types of facilities. The low scores on examination resulted mainly from patients not being examined at all. Regarding treatment, private clinics and pharmacies scored low because often only one causative agent of an STD syndrome was treated in the absence of laboratory results. This type of management, common before syndromic management was introduced, apparently is still widespread in Nairobi despite its lower effectiveness. ^28–30

Scores for history taking and examination were highest during the interview, lower during the observation, and lowest during the simulated patient visit. This implies that providers have a fairly good knowledge of these aspects, but that they are not translating it into daily practice. A possible reason for this is the high workload some providers are facing. For treatment, pharmacies and private clinics scored below 40% even during the interview, which indicates that knowledge of syndromic treatment is highly insufficient in these facilities.

The overall score for STD case management as measured by PI6 varied from 14% to 48% during patient observations, and from 11% to 31% during simulated patient visits. To be fully cured and not immediately reinfected, correctly managed patients must comply with the prescribed medication, come back to receive alternative treatment if the problem persists, have their sexual partners treated in a correct way, and use condoms or abstain from sex during the course of their own and their partner’s treatment. Therefore, the proportion of patients eventually cured will be alarmingly marginal. This shows that aspects of health education such as condom use, contact-tracing, compliance, and counseling are crucial in the prevention and control of STD/HIV. These aspects also were assessed in the current study, but are reported separately. ²³

The study findings indicate the need for action to improve STD care in the private sector of Nairobi. Because knowledge is insufficient, using private clinics and pharmacies to train providers in syndromic management is highly recommended. Several studies have proved that training can significantly improve the performance of providers in STD management, at least in the short term. ^17,19,31 To sustain provider behavior change, training should be an ongoing process with follow-up evaluation and refresher courses. ¹⁷ In this study, the differences found between providers trained and those untrained in STD management proves that training indeed has a positive impact on both the knowledge and performance of providers. This finding confirms the importance of further provider training in Nairobi.

Alternative interventions are needed to complement training, especially to improve history-taking and examination. Several studies have shown that inadequate supervision is one of the factors contributing to low quality of STD care. ^32,33 Considering that in Guinea-Conakry PI6 increased from 7% to 48% within 2 years by improving supervision, this might be an effective intervention strategy in Nairobi as well. ³⁴ An additional option, especially for pharmacies and the private sector, is the development and social marketing of prepackaged syndromic management kits for STD. These kits, which contain drugs for one STD syndrome, condoms, and partner notification cards, have proved to be feasible and acceptable for health workers and patients in several pilot studies. ^35–37 Health workers believed the packages improved STD management by making treatment easier and saving time. ³⁵ In a case–control study, patients using these kits had significantly higher self-reported cure rates than their controls (84% versus 47%). ³⁶

To which facilities can interventions best be targeted? Because facilities differ most significantly from each other in terms of treatment, interventions can best be targeted to pharmacies and private clinics, which score lowest on treatment. For interventions to be cost effective, the STD caseload per provider also must be taken into account. Because pharmacies have a higher daily STD caseload per provider than private facilities, as determined by the number of patient observations, it could be especially cost effective to target interventions to pharmacies. Training in syndromic management and the introduction of prepackaged STD syndromic management kits would be important interventions for pharmacies because they do not perform examinations or laboratory tests. For effective implementation of any intervention targeted to this group, it also will be crucial to have the support of organizations such as the Kenyan pharmacy association.

Improved STD care may eventually lead to improved healthcare-seeking behavior of patients. A complementary approach to improve STD control focuses interventions directly on the patient’s health-seeking behavior. If more patients with STDs seek care at facilities with high-quality care, the overall cure rate of STDs improves. For the Nairobi situation, this means that patients with STDs should be convinced to go to public clinics equipped for STD care. However, patients generally have the idea that quality of care is highest in private facilities, which are more expensive than public facilities. Therefore, Information, Education and Communication (IEC) activities should focus on raising quality awareness among patients with STDs and on changing their health-seeking behavior. These interventions should go hand in hand with interventions focusing on healthcare providers to improve STD case management.

Although it is difficult to compare the quality of STD case management across countries, ³⁸ an attempt is made to compare the Nairobi study with studies using similar outcome measures. A study in Malawi, where public and private practitioners were observed, found the same rate for history-taking as in Nairobi (81%), but lower rates for examination (46% versus 51%), treatment (13% versus 54%), and PI6 (11% versus 27%). ⁶ A study in South Africa among public sector providers using simulated patients found a score of 45% for correct history-taking (75% to 100% in Nairobi public clinics), 19% for examination offered (44% to 50% in Nairobi), and 41% for correct treatment (75% in Nairobi). ⁷ Studies in Peru and Nepal each found the percentage of simulated patients who received correct syndromic treatment in pharmacies to be only 1% (45% in Nairobi). ^16,17 Findings showed that PI6 scores based on observations were 12% in the public and private sectors in India (27% in Nairobi) and 4% in public clinics in Ethiopia (15% to 48% in Nairobi). These comparisons indicate that the quality of STD case management in Nairobi generally is better than in other studied developing countries.

In retrospect, this study could have been improved in several ways. We mention these so others can build on our experience. First, more providers could have been studied by all three methods if the identity of the providers visited by the simulated patients had been monitored. If the simulated patient had asked the provider’s name, then the interviewer/observer could have made an appointment with the facility on a day when this specific provider was at work. Second, the gap between knowledge and practice could have been anticipated by explicitly asking providers in the interview to identify barriers to translating knowledge into practice. This would have given fruitful suggestions for the development of strategies to address the gap between knowledge and practice. Third, the power of the comparative analyses could have been increased if resources had allowed us to sample more than five study areas.

In conclusion, although STD case management in Nairobi measures up to that of other developing countries, only 27% of the observed patients with STDs are being managed correctly. This indicates that the overall cure rate of patients with STDs in Nairobi can be increased. Recommended interventions are ongoing training in syndromic management in the private sector as well as improved supervision and the introduction of prepackaged syndromic management kits.

Acknowledgments

The authors thank the heads and staff of all studied facilities as well as the patients with STD for their participation and cooperation. They also thank the Ministry of Health, the district commissioners, and (assistant) chiefs in the study areas for their continuous support. Furthermore, they express their appreciation for the dedicated work of their research assistants, who worked tirelessly in collecting data.