Operations Research

A job aid shows a mother giving a child medication with a spoon. The father is also involved. Artist: Mahamadou Nadéré

Client Satisfaction

Client satisfaction in Niger
This study examined the utility, validity, feasibility, and cost of several types of client satisfaction data collection methods in three districts of Niger. Patient satisfaction data were collected from exit interviews with 301 clients and from 24 client focus groups. Three types of data collectors were used: health supervisors from the same district, outside enumerators, and health supervisors from a neighboring district. The advantages and disadvantages of the two data collection methods and three data collector types were measured using key informant interviews with district managers, questionnaires to the data collection teams, cost records, and analysis of the exit interviews and focus group data. The study found that exit interviews were more valid and useful than focus group data and were also less costly in terms of data collector time. Supervisors from the same district were less costly, more useful, but slightly less valid than the alternatives. Outside enumerators may be more feasible by not drawing on overworked supervisors.

Download report:
Helping District Teams Measure and Act on Client Satisfaction Data in Niger

Measuring client satisfaction at MaxSalud clinics in Peru
This study examined the advantages and disadvantages of six different methods for obtaining client satisfaction information at the MaxSalud clinics in Peru. The methods included: client exit interviews (323 clients), client home follow-up visits (32 visits), focus groups (8), home interviews with discontinued clients (40 visits), suggestion boxes, and community meetings. Exit interviews were relatively inexpensive, quantitative, periodic, and useful for ongoing monitoring. Follow-up visits provided valid, in-depth information useful for quality assurance work. Focus groups were useful but not very feasible. Discontinued client interviews were less useful and relatively expensive. Suggestion boxes are easy to establish but don't produce much useful information. Community meetings provide contextual information, not useful client satisfaction data. Clinic managers and quality committees appreciated and used the client satisfaction data, especially when it was succinctly summarized.

Download report:
Implementing a Client Feedback System to Improve the Quality of NGO Healthcare Services in Peru

Activity-based costing at MaxSalud clinics in Peru
Activity-based costing (ABC) allocates indirect costs to outputs according to the actual work done to produce the outputs rather than in proportion to the relative volume or direct costs of outputs as in traditional cost accounting. This study applied ABC to calculate the unit cost of services provided by MaxSalud clinics in Peru. The application included: (1) the description of all departments, services, and activities by department, (2) staff estimates of time spent on each activity and unproductive time, (3) estimated cost of all activities by each department using wage and other data, (4) trace activities and costs within and across departments to services provided, and (5) estimated service volumes from records and calculate unit costs (cost/volume). The study identified 107 distinct activities at MaxSalud, including training and meetings. ABC derived unit costs that were generally higher than prior estimates and much higher than fees charged. The study concluded that ABC is potentially very valuable to MaxSalud to set policy, manage expenditures, and even raise funds, but it requires reliable data systems for costs and service statistics, management attention, staff support, and technical assistance.

Download report:
Application of Activity-Based Costing (ABC) in a Peruvian Based NGO System

Compliance, workload, and cost of IMCI in Niger
We studied the relationship between IMCI compliance, cost (drugs and provider time), and workload for 211 child cases of fever, cough, diarrhea, and earache in 26 health clinics in Niger. IMCI compliance rates were 33 percent for assessment, 81 percent for treatment, and 42 percent for counseling. Key result: no relationship was found between compliance and cost, even for particular diagnoses, nor between average clinic compliance and average clinic workload. This result is interesting because previously published studies have reported that increased compliance with IMCI guidelines have resulted in lower drug costs; this study suggests this is not always true. On the other hand, this study found that increased compliance did not lead to increased costs, either for drugs or for consultation time.

Download report:
Compliance, Workload, and the Cost of Using the Integrated Management of Childhood Illness Algorithm in Niger

Impact of new guidelines for arterial hypertension care on costs in Tula Oblast, Russia
A quality assurance project in Tula Oblast, Russia introduced new guidelines for the management of arterial hypertension (AH) in 1999 that emphasized screening, patient education and control of blood pressure. This OR study tested the hypothesis that the new guidelines would shift resources from expensive inpatient care to less expensive outpatient care, thereby resulting in a net reduction in the cost of AH care. An analysis of inpatient and outpatient costs associated with AH before and after the new guidelines in a population of about 10,000 adults supported the hypothesis. AH inpatient admissions dropped 17%, patients under outpatient AH care increased 47%, the cost per 1000 adults of inpatient AH care dropped 32%, and the cost of outpatient AH care per 1000 adults rose 61%. Because inpatient care is far more expensive than outpatient care, overall cost of AH care per 1000 adults dropped 11%. Most of the reduction in inpatient admissions and cost was due to a sharp reduction in the number and cost of emergency admissions for AH care by patients not previously under outpatient care for AH. After reviewing limitations of the study methodology and possible alternative causes, the study concluded that the drop in AH cost was almost certainly due to the new guidelines.

Download report:
Assessing the Economic Impact of the New System of Care for Arterial Hypertension in Tula Oblast, Russia

Improved quality of care reduces pregnancy-induced hypertension costs in Tver Oblast, Russia
QAP helped Tver Oblast redesign its system of care for pregnancy-induced hypertension (PIH) in three hospitals, including the introduction of new, evidence-based guidelines. Under the new guidelines, fewer women were admitted for PIH care, and those admitted received more aggressive care to prevent progression to eclampsia, including mono-drug therapy (magnesium sulfate) and early induced delivery if the drug therapy did not work. The new system resulted in no progressions to eclampsia, no PIH deaths, and a 60% drop in newborn complications to PIH mothers in the 19 months following the introduction of the new guidelines. This study focused on measuring the cost of the new and old systems in two of the hospitals. The direct cost of drugs, laboratory tests and hospitalization (nurse and clinician time, food) for PIH inpatient and outpatient care was estimated for PIH patients in similar 6-month periods before and after the new guidelines. Total costs dropped by 87% following the introduction of the new guidelines. Four factors contributed most to this reduction: a 76% drop in the number of cases admitted, a 13% reduction in the length of hospitalization, an 81% reduction in the number of women given additional PIH outpatient care over and above normal antenatal care, and more rational use of drugs, resulting in a 63% reduction in the drug cost per inpatient.

Download report:
Improving the Quality of Care for Women with Pregnancy-Induced Hypertension Reduces Costs in Tver, Russia

Measuring the cost of inefficiency in hospital labs in Ecuador
This study developed a method for identifying inefficiencies in hospital laboratories and measuring their cost, and tested it in three hospitals in Ecuador. Seven types of inefficiency are included: (1) ordering unneeded tests, (2) test results never picked up, (3) inefficient use of lab resources, (4) low productivity of laboratory staff, (5) expiration of lab reagents, (6) lack of quality and cost control practices, and (7) inefficient procurement of lab reagents and other resources. Theft and long-term costs of poor quality were not addressed. It was necessary for the hospitals to develop explicit standards before the cost of unneeded tests could be measured. Data collection methods included review of existing lab records, financial records, hospital utilization records, and a sample of medical records; observation of high frequency tracer tests to estimate time used; and key informant interviews. The study found that staff time was the largest cost factor and that under-utilization of staff is a potentially large component of inefficiency. Expired reagents were a significant contributor to inefficiency in all three hospitals. High costs of resources due to poor procurement practices were a major source of inefficiency in at least one hospital. Unneeded tests comprised about 7% of all costs due to inefficiency, while three types of inefficiency were not important (tests not picked up, inefficient use of lab resources, lack of practices).

Download report:
Measuring the Cost of Inefficient Use of Laboratory Resources: Ecuador

Effectiveness of QA Interventions

Assessment and improvement of family planning supervision in Zimbabwe
To assess the quality of supervision at the district level, data were collected from various sources: structured observations of supervisors, audiotaping of supervisor-provider interactions, recording of all supervisory activities, and interviews with supervisors and supervisees. A team composed of current and past supervisors, along with researchers, determined the supervisory practices that would be measured. Relying on supervision standards developed by Zimbabwean stakeholders, the study found zero percent of supervisors performed "high," 56% performed "medium," and 44 percent performed "low." Supervisors devoted less than 5 percent of their time to patient care issues. The supervisors' main strengths were in giving feedback on technical standards, discussing and analyzing data, and developing rapport with the providers. They were most deficient in making suggestions, seeking client input, problem solving with the providers, and building on previous (and future) supervisory visits. The supervisors' weakest areas were continuity of supervisory visits, being proactive, and engaging in joint problem solving with front-line providers. Building on this formative study, QAP designed an intervention to enhance supervisory skills and improve patient care using supervisor led, on-job training of providers, and included a detailed course curriculum. The course was implemented in Zimbabwe's Mashonaland East Province, and the curriculum revised and published.

Download report:
The Quality of Supervisor-Provider Interactions in Zimbabwe

Evaluation of computer-based training for IMCI in Uganda
This study compared the cost-effectiveness of a 9-day IMCI computer-based training course and an 11-day standard classroom IMCI training course on the knowledge, skills and performance of providers. The standard course was given to 55 providers, and 59 took the computer-based course. Tests of IMCI knowledge were given to all course participants before the training, immediately after the training, and 3-4 months after the training. Skill and performance in case management were observed by trained observers immediately after the training and again after 3-4 months. Both types of training led to led to significant and sustained improvements in participants' knowledge of IMCI, but there was no difference between the two training methods. No differences in skills were noted between the two types of training at the end of the course or after 3-4 months, although there was a substantial decay in skills over the 3-4 month period between observations. Because the computer course took less time and required fewer facilitators to administer, it cost 29% less than the standard course per trainee$335 per trainee for the computer course versus $472 per trainee for the standard course.

Download report:
A Comparison of Computer-Based and Standard Training in the Integrated Management of Childhood Illness in Uganda

Factors influencing success of quality improvement teams in Morocco
A sample of 23 quality improvement (QI) teams from 27 health facilities (9 hospitals, 18 health centers) in four regions of Morocco were studied to determine the effect of coaching and functionality on their success. Functionality was defined as the degree to which the team followed a simplified 4-step QI problem-solving protocol (identify problems, analyze and select the key problem, develop and implement solution, and define and measure monitoring indicators). Data were obtained about the number of problems defined, problem-cycles completed, and success in implementation and problem solving through structured questionnaires and interviews with team members, team coaches, and regional QA directors. Success for a particular problem-solving cycle was defined by whether the respondents said the team accomplished or not the chosen objective and by whether the monitoring indicator showed improvement, no change, or decline. With respect to functionality, teams did well on problem identification (80%) and problem analysis (90%), but poorly on solution development and implementation (55%). Of the 14 teams that completed at least one problem-solving cycle, 67% engaged in good monitoring techniques, and 80% reported achieving improvement in their indicator of interest. This suggests that functionality is related to success as measured.

Download report:
Quality Improvement Teams in Morocco: An Evaluation of Functionality and Success

Download report:
Impact of QA Methods on Compliance with the Integrated Management of Childhood Illness Algorithm in Niger

Impact of self-assessment and peer feedback on health worker IMCI performance in Mali
This study examined the effect of self-assessment and peer feedback on the quality of care provided to children with fever in a peri-urban area of Mali. Quality was defined as the percentage of assessment and counseling tasks performed according to nationally accepted standards. A sample of 36 providers was randomly assigned to a program group (12) and a control group (24), and then observed managing 103 cases of child fever (about 3 cases per provider). The program group received the self-assessment and peer feedback interventions, while the control group did not. In the program group, 54% of the tasks were performed according to standard, compared to only 44% in the control, a statistically significant difference. However, users found the self-assessment to be time-consuming and burdensome.

Download report:
The Impact of Self-Assessment with Peer Feedback on Health Provider Performance in Mali

QA teams and IMCI compliance in Kenya
This study tested whether facility-based QA teams, trained and coached to develop and implement solutions to IMCI problems, could improve IMCI case management after one year. The study looked at 21 QA teams that implemented a variety of improvements: procure more IMCI drugs, clocking-in register, on-the-job IMCI training, patient IMCI education, workload sharing by staff, reduce waiting time, and monthly meetings. Approximately 70 providers were observed attending about 10 IMCI cases each before the team interventions, and again after the interventions, in the 21 facilities with QA teams and in 14 facilities without QA teams. Case management was scored against a set of indicators for IMCI assessment, classification, treatment, and counseling, and change in performance was analyzed for the 59 providers observed both before and after. Quality of case management improved by 57 percentage points in the facilities with QA teams, compared to only 14 percentage point improvement in the control facilities.

Supportive supervision and self-assessment in Mexico
This study introduced a program of supportive supervision and physician self-assessment to improve provider-client communication with 60 doctors working for Mexico's Social Security hospitals. The intervention was evaluated using audio-taped counseling sessions obtained pre and post in a program and a control group. The study found that facilitative and informative communications per counseling session by the doctors in the program group were approximately twice that in the control group following the intervention; the difference between program and control was statistically significant even after controlling for purpose of visit, sex of doctor, and duration of session. In a pre-post panel analysis of 28 doctors, substantial improvements occurred in both the program and control groups, but the increases were markedly larger in the program group, where increases were about twice as great in provider facilitative and informational communications. Active patient communication, as measured by the number of questions asked by the client, increased dramatically and by the same amount in both the program and control group (from 2.4 to 12.7 questions per session), a finding consistent with the post-intervention result for 60 doctors where no significant difference was observed in active patient communication. The fact that session duration increased pre to post but was not significantly different in program and control indicates that the rate of provider communication went up in the program group. The self-assessment experience by the supervisors and the doctors was highly regarded by them, as was use of tape recorders by the doctors to listen to their own counseling sessions.

Download report:
Participatory Supervision with Provider Self-Assessment Improves Doctor-Patient Communication in Rural Mexico

Evaluation of QA Programs

Evaluation of the Chile National Quality Assurance Program
The Chile National Quality Assurance (QA) Program began in 1991, and working through a QA Unit in the Ministry of Health, developed standards, training modules, and communication strategies, eventually achieving country-wide coverage. This evaluation was undertaken in 1999 by a team of three senior staff from QAP and the director of the national program to develop and apply an evaluation methodology that would document the QA program's structure and management and its technical and support functions. The methodology included: (1) developing a logical framework of technical areas, (2) identifying questions for each technical area, and indicators, data sources, and data collection methods for each question, (3) setting the timing for data collection (before or during the evaluation team visit), and (4) analysis and write-up of the findings. The evaluation revealed a very successful program with many innovative characteristics. About 12,000 health professionals had been trained in QA methods, representing 25 % of the health workforce. The Chilean QA program is closely linked to primary care and public health, with QA more prevalent in health centers than larger facilities. Incentives have been in the form of professional awards and public acknowledgement. The team concluded that strong dedication to QA by health professionals throughout the country has developed close bonds at the local level that have been crucial to QA's success there.

Download report:
Evaluation of the Chile National Quality Assurance Program

Evaluation of the Zambia Quality Assurance Program
In 1998, the Central Board of Health (CBoH) asked QAP to review the performance of the Zambia Quality Assurance Program and provide recommendations to help design its next phase. QAP staff led an international team of six experts from the United States, Norway, Denmark, and Zambia to conduct the quality assurance program evaluation, which focused on: (a) development and communication of standards of care and measurement of compliance with those standards, (b) productivity of quality improvement teams using a problem-solving method, and (c) the existence and effectiveness of the support systems for QA activities. Topics for evaluation were identified from a systems analysis of what an "ideal" quality assurance (QA) program might look like. The evaluation team selected nine districts for the field interviews, purposely choosing districts with health centers that had active QA teams or staff who had received some degree of QA training. In these nine districts, representing all four regions of Zambia, 24 health facilities were visited, and a total of 140 persons were interviewed.

Download report:
The Zambia Quality Assurance Program Final Evaluation

The Niger QAP/BASICS Joint Project Final Evaluation
QAP launched a pilot program in Niger in 1993 to strengthen newly decentralized regional and district management and improve the quality of primary healthcare in the regional Department of Tahoua. The project sought to develop the capacity of regional and district management teams to apply quality management (QM) techniques to improve the quality and client-focus of basic services, strengthen teamwork, and improve the measurement and monitoring of results. The Basic Support for Institutionalizing Child Survival Project (BASICS) became active in Niger in 1995, helping to improve child health services, especially the care of sick children. In 1997, the two projects merged in Niger to form a joint project, QAP/BASICS. The final evaluation of the joint project was conducted in 1998 and had two objectives: 1) to assess the degree of quality assurance institutionalization in the Tahoua Department as a result of the work of the Tahoua Quality Assurance Project since 1993, and 2) to summarize lessons from introducing the Integrated Management of Childhood Illness (IMCI) guidelines in a quality management (QM) environment. The evaluation team collected data from the two departments of Niger in which the joint project was carried out: Tahoua and Dosso. The team visited the two departmental headquarters, six district headquarters, and 13 integrated health centers. These represent 80% of the facilities involved in the project. Using structured questionnaires, the team interviewed approximately 60 staff members, covering all levels of the healthcare system in both departments, as well as international partners (BASICS, WHO, UNICEF, and German Cooperation representatives). To measure IMCI performance, the team reviewed a sample of completed IMCI forms in three facilities each in Boboye and Konni Districts. The evaluation team also organized two focus groups of women who had used prenatal care and child health services in Tahoua. The team prepared three case studies to document quality improvement processes and results in three facilities.

Download report:
The Niger QAP/BASICS Joint Project Final Evaluation

The sustainability of quality assurance activities in Tahoua, Niger
This study assessed the continuation of quality assurance activities in the Tahoua region of Niger two years after the close of the QAP/BASICS project. The project, which operated from 1993 to 1998, provided technical and operational support to basic health services in order to improve their quality through training, supervision, meetings, operational research and provision of small equipment. This study compared current QA activities in 8 districts of Tahoua and 4 districts of a control region (Maradi) during the two years (1999 and 2000) after the QAP/BASICS project closed. Interviews of Niger Ministry of Public Health personnel and review of documents were conducted at all levels in the health system. Healthcare workers were observed to determine compliance with IMCI standards and caretakers were interviewed after the observation. In the last two years in Tahoua, more standards were revised than in Maradi (4 vs. 2), and more district coordination meetings occurred in 2000 (44 vs. 32) but not in 1999 (56 vs. 58). The percent of planned supervisory visits actually completed was higher in Tahoua than in Maradi in 1999 (51 vs. 24) but not in 2000 (52 vs. 55). The number of persons trained in QA continued at approximately the same level after the close of the project (22 in 1997, 79 in 1998, 26 in 1999, and 76 in 2000), but the number of problem-solving cycles completed by QA teams decreased after project close (31 in 1998, 21 in 1999, and 17 in 2000). Several factors helped maintain the project's accomplishments: new funding, availability of trained staff, demonstrated results, and staff commitment. However, the regularity of supervision and the QA team productivity suffered after the project closed.

Job Aids

Case management maps for pregnancy-induced hypertension and postpartum hemorrhage
This study developed, implemented and measured the impact of a case management map on the management of pregnancy-induced hypertension (PIH) and postpartum hemorrhage (PPH) in Jinja Hospital, Uganda. The maps served as a job aid and a permanent record of the care provided. The map development process within the hospital was an integral part of the intervention, including the steps of team selection, current process description, developing the job aid map, define monitoring indicators, and implementation. The development team also worked to solve barriers to good performance such as missing equipment or drugs. This process was undertaken twice in the same ward, first for PIH and six months later for PPH. A third maternal condition, pelvic inflammatory disease (PID), was used as a control. Impact was measured by quality of care and by health outcome. To measure quality of care, three key tasks (one each for assessment, monitoring and treatment) were measured for six months pre-intervention and six months post intervention for all three conditions. The average of the three quality tasks performed to standard in 86 PIH cases increased from 0.33 to 2.3 (significant), and the average for 30 PPH cases increased from 0.45 to 1.1 (not significant), while the performance in 66 PID cases increased only slightly (0.46 pre, 0.62 post, not significant). Health outcomes improved for PIH cases (stillbirths and women progressing to eclampsia both dropped but not significantly), but did not improve for PPH cases (maternal deaths increased). The study concluded that the case management maps contributed to increased quality of care under ideal conditions.

Job aids for IMCI in Zambia
This study investigated how a job aid could be used to increase compliance with IMCI guidelines in Zambia. Prior to the study, Zambia had trained many providers in IMCI, and several IMCI job aids were in use. The study included surveys of job aid usage and IMCI compliance in 1999 and in 2000, and the design and introduction of a new IMCI job aid between the two surveys. The 1999 survey observed 385 IMCI cases by 57 providers in 33 health centers. The 2000 survey observed 263 IMCI cases by 56 providers in 16 health centers, and reviewed a random sample of 157 completed IMCI records in the new job aid format. The surveys also interviewed providers and caretakers. Both surveys found high correlation between the use of job aids for IMCI case management and compliance with IMCI standards. This result does not demonstrate that job aids cause improved compliance, perhaps because the study methodology did not address whether individual providers who previously did not use a job aid improved their compliance when they started using job aids. There was no significant difference in compliance between users of the new job aid and the existing IMCI chartbook in 2000. Most providers (82%) said the new job aid was useful because it saves time, reduces errors, helps them remember the IMCI algorithm, and is easy to use, although they also used the existing IMCI chartbook frequently and for the same reasons. Nearly all caretakers said they preferred providers to refer to written material such as a job aid during the IMCI session, thus challenging the contention by some providers that using a job aid would make them look bad in the eyes of patients.

Download report:
Assessing the Functionality of Job Aids in Supporting the Performance of IMCI Providers in Zambia

Job aids to improve adherence to cotrimoxazole treatment in Niger
This study tested the effect of a set of job aids and provider training on provider, caretaker, and patient adherence to cotrimoxazole for the treatment of childhood pneumonia in Niger. Data on 677 cases of childhood pneumonia were obtained from home and clinic observations and interviews in 4 experimental and 4 control clinics following the introduction of the job aids in the experimental clinics. The study found that caretaker adherence to the recommended regimen was much higher in the experimental than control clinics for cases treated by technical nurses (lower educational level), while little effect was observed on professional nurses (higher educational level). In fact, the performance of technical nurses in the experimental clinics rose to near the performance of the professional nurses, suggesting that the job aids may be a low-cost alternative to more expensive formal education. Another interesting result was that providers should prescribe and give caretakers the full course of pills at the initial visit. Partway through the study it was discovered that the clinics were giving a partial course of cotrimoxazole at the initial visit (1 to 3 days of pills rather than 5 days) for a variety of stated reasons (e.g., force caretaker to return for follow-up visit, misunderstood recommended practice, low inventory of pills). In fact, the study showed that children were far more likely to consume the full course if their caretaker received it at the initial visit than if not. This study is relevant to curbing antimicrobial resistance.

Download report:
Improving Adherence to Cotrimoxazole for the Treatment of Childhood Pneumonia in Niger

Job Aids Symposium
The Quality Assurance Project and Child Survival Collaborations and Resource (CORE) Group cosponsored a symposium to exchange views and evidence about the state-of-the-art in job aids, to share particular job aids that have been successfully used in international health, and to identify future developments that will make job aids more useful in field applications for child survival and international health. Seminar participants heard from leading practitioners in the field of job aids on the evidence of their effectiveness and practical applications of job aids for health in developing countries.

View and download presentations and transcripts from the symposium

Malaria rapid diagnostic kits
As a complement to a CDC study in Malawi, QAP developed and implemented a research protocol for testing and modifying job aids (instructional inserts) that accompanied two Malaria Rapid Diagnostic Kits. The study used quality design techniques to improve the job aids and to make recommendations for product design changes. The technique involved observing kits being used, noting problems, redesigning the instructions, having users test them while under observation, and redesigning them again. In an initial test prior to the modification of the instructional inserts, only 3 of 19 (15%) providers used the two kits without making errors. Eight of the providers had received training in the use of the kits, and two of these (25%) were error free. After the instructions were revised, 85% (17 of 20) used the kits without error. The study demonstrated that large improvements can be achieved in the instructions by using the redesign process and that improved inserts can be more effective than training.

Download report:
Using Quality Design to Improve Malaria Rapid Diagnostic Tests in Malawi

Vendor-to-vendor education to improve malaria treatment by private drug outlets
Although many patients obtain anti-malarial drugs from small private drug outlets in Kenya, earlier studies indicated that 87% of shopkeepers had never received training in appropriate use of anti-malarial drugs and that 60% gave instructions or dosages to customers. This study tested whether a low-cost education strategy in one district of Kenya would increase private drug outlet knowledge of and compliance with national malaria guidelines. Local wholesalers (mobile vendors and counter attendants at wholesale outlets) were given a one-day training in malaria drug standards and equipped with customized job aids (posters) to give to the retail drug outlets that were their customers. Mystery shoppers visited 101 intervention outlets that received the job aids from wholesalers and 151 control outlets that did not receive the job aids, asking for treatment for their child under two distinct scenarios that required decision-making by the shopkeeper. In response, the mystery shoppers were sold over 70 different anti-malarials (only 5 are government approved), and over 30 anti-pyretics. In the intervention outlets, the shoppers came away with an approved anti-malarial and were told the correct dosage 17% (27/157) of the time compared to 1.5% (3/202) in the control outlets. In the intervention outlets, 35% of shopkeepers answered all 10 knowledge questions correctly compared to only 4% in the controls. The intervention cost was about $17 per outlet reached, or about $0.10 per additional case correctly treated.

Download report:
Vendor to Vendor Education to Improve Malaria Treatment by Drug Outlets in Kenya

Regulatory Strategies
Development of the national hospital accreditation program in Zambia
This study describes the development of the Zambian national hospital accreditation program from 1997 to 2000. Data were collected through a review of written documents, interviews with major stakeholders, hospital visits, and discussions with implementers. Zambia successfully developed hospital standards that are relevant and potentially achievable by its hospitals. Half of Zambia's 79 hospitals have received educational surveys, and twelve have also received the full accreditation survey. Significant improvement in compliance with standards occurred in overall scores and in 7 of 13 functional areas. However, the program has stalled due to lack of sufficient funds, lack of legal standing for the Zambia Hospital Accreditation Council, difficulties in retaining qualified surveyors, and indecision on how to handle accreditation results. In addition, hospitals' serious resource constraints and need for ongoing facilitation have hindered their full participation in the program. It is estimated that the program costs about US$ 10,000 per hospital to complete the cycle.

Download report:
Setting Up a National Hospital Accreditation Program: The Zambian Experience

Impact of hospital accreditation in KwaZulu-Natal, South Africa
KwaZulu-Natal (KZN) province of South Africa contracted with the Council for Health Services Accreditation of Southern Africa (COHSASA) to introduce hospital accreditation into their public hospitals in 1998. QAP undertook a study of the impact of the COHSASA program in 20 randomly selected hospitals, stratified for size, in KZN, with ten serving as program hospitals and ten as controls. The study analyzed the effect on the standard COHSASA indicators measured as a normal part of their accreditation process, and on a limited set of quality indicators developed by a consensus process and collected by an independent research team. The quality indicators were: nurse perceptions of quality, client satisfaction, client medication education, accessibility and completeness of medical records, quality of peri-operative notes, hospital sanitation, and labeling of ward stocks. Unfortunately, several indicators more directly related to clinical outcomes were dropped due to lack of data. About two years after initiation of the program, the program hospitals had improved their average compliance with COHSASA accreditation standards from 38% to 76%, while no increase was observed in the control hospitals (from 37% to 38%). This improvement of the program hospitals relative to the controls was statistically significant and seems likely to have been due to the accreditation program. However, little or no effect of the program was observed on the indicators collected by the independent research team, with the exception of nurse perceptions of clinical quality. Limitations of the study design may have influenced the observed results. Practical implications of these results are: (1) the COHSASA facilitated accreditation program appears to be successful in improving hospital performance on COHSASA standards, and (2) additional work is needed to determine whether improvements in COHSASA standards of structure and process result in improved quality of care measures.

Impact of the hospital accreditation program in Zambia
This study assessed the impact of Zambia's nascent hospital accreditation program on the quality of care. Twenty hospitals just starting in the accreditation program formed the study program group, and six non-participating hospitals formed the control group. Data were obtained from the accreditation surveys on hospital performance against the accreditation standards, and from additional research surveys on process and outcome indicators. The study found that hospitals exposed to the accreditation program seemed to have achieved higher performance on standards than the unexposed. The accreditation program has had the greatest impact on general hospitals, followed by district, and lastly mission hospitals. Compliance with accreditation standards appeared linked to improvement in some, but not all, indicators of quality. Management support and good information sharing seemed to lead to higher compliance with accreditation standards.