OUP user menu

Hospital quality improvement in Ethiopia: a partnership–mentoring model

Elizabeth Bradley, Kari A. Hartwig, Laura A. Rowe, Emily J. Cherlin, Josh Pashman, Rex Wong, Tim Dentry, W. Edward Wood, Yigeremu Abebe
DOI: http://dx.doi.org/10.1093/intqhc/mzn042 392-399 First published online: 10 September 2008


Background and Objective Quality improvement efforts are increasingly common in the United States; however, their use in developing countries is limited. We sought to evaluate the impact of a large-scale intervention on several key management indicators through hospital quality improvement efforts.

Design Pre–post-descriptive study of 14 hospitals in Ethiopia.

Setting Six regions and two city administrations in Ethiopia.

Participants Hospital leaders and management mentors in participating hospitals.

Intervention In collaboration with the Ministry of Health and the Clinton HIV/AIDS Initiative, we implemented a countrywide quality improvement initiative in which 24 mentors with hospital administration experience were placed for 1 year in Ethiopia to work side-by-side with hospital management teams. We also provided a professional development course to enhance quality improvement skills.

Main Outcome Measure(s) Presence of 75 key management indicators; reported management skills of hospital leaders by the mentors.

Results In pre–post analysis, we found improvement in 45 of the 75 (60%) key management indicators between August 2006 and May 2007. The changes reflected a total of 105 management indicators improved across the 14 hospitals, which equates to a per-hospital mean of 7.5 (standard deviation 5.9) improvements. Reported management skills of hospital leaders improved in several management domains, although their reported confidence in these skills remained largely unchanged.

Conclusions Our findings indicate that quality improvement efforts can be effective in improving hospital management in developing countries. Longer follow-up is required to assess the sustainability of the hospital improvements accomplished.

  • Ethiopia
  • quality improvement
  • hospital management
  • Africa


Quality improvement efforts to promote greater management effectiveness within hospitals are increasingly common in the USA [13]; however, use of quality improvement effort in developing countries is modest. Recent ‘World Health Reports’ published by the World Health Organization have called for health system strengthening as paramount to the goal of improving health outcomes [4, 5], noting the critical role that hospitals play in efforts to strengthen health systems. Despite the potential role of quality improvement in health system strengthening [69], we have only limited evidence about its implementation and effectiveness in developing countries.

Studies of healthcare quality improvement in developing countries suggest its positive impact on quality-related outcomes, although the scope of the studies is narrow [1015]. Three studies have been conducted in single hospital settings [10, 12, 13], and two additional studies [11, 15] have been conducted in primarily health center settings. Using pre–post observational designs, the studies uniformly report improvements in care as measured by specific indicators such as availability and cleanliness of latrines, use of stock cards for pharmacies, improved floor washing and availability of adequate equipment [11, 14], and in perceptions of more positive organizational culture and overall enhanced quality of service [10, 12, 13, 15]. Despite the consistency of these findings, however, the evidence about the impact of quality improvement efforts in developing countries remains limited because studies typically include one or a handful of sites with little quantitative data to evaluate the impact of the interventions on targeted outcomes.

In this study, we sought to examine the impact of a large-scale intervention, the Ethiopia Hospital Management Initiative (EHMI), to improve hospital management through the use of quality improvement methods in hospitals. As has been previously described [16], the EHMI employs a partnership–mentoring model, which incorporates the principles and tools of quality improvement including participatory approaches to organizational change. Under the direction of the Ethiopian Minister of Health, EHMI was launched in January 2005 in partnership with the Clinton Foundation HIV/AIDS Initiative (CHAI) and the Yale School of Public Health. During 2005–06, EHMI was piloted in 14 government hospitals across Ethiopia with the goal of replicating successes from the pilot efforts in a broader set of hospitals throughout the country. Understanding the impact of this large-scale intervention can help further assess the potential of quality improvement in strengthening health systems in developing countries.



Since 1994, Ethiopia has been managing a decentralized health system across nine regional states and two city administrations (Addis Ababa and Dire Dawa). The Federal Ministry of Health (FMOH) develops policy and technical guidelines, which are managed by Regional or City Administration Health Bureaus (RHB). Hospitals remain a primary point of access for healthcare services within the Ethiopian health system. Although there has been a significant increase in the number of private healthcare facilities in largely urban centers throughout the country, the majority of the population depends on government hospitals for their health care.

In this study, we implemented pilot quality improvement projects in 14 government hospitals within six regions and two city administrations. Three of the hospitals were located in rural settings, four were located in semi-urban settings and seven were located in Addis Ababa. The 14 hospitals included one teaching hospital and a mix of specialized, referral and district hospitals with number of beds ranging from 74 to 560, with a mean of 240 beds. The general manager at each hospital was a medical director (always a physician). With the exception of a financial administrator, all members of the management team were clinicians (e.g., doctors, nurses and pharmacists). At baseline, the primary challenges in these hospitals included low staff morale; shortage of financial resources, medical equipment, drugs and other supplies; limited opportunities for staff training; poor infection control and a lack of community participation in hospital governance [16].

Study design

We conducted a pre–post study of the 14 hospitals that participated in the EHMI, under the leadership of the Minister of Health and in partnership with CHAI. A detailed description of the partnership–mentoring model used in each hospital is elaborated elsewhere [16] and described briefly here. The Yale University team recruited 24 Senior Yale–Clinton Foundation Fellows and Post-Graduate Yale–Clinton Foundation Fellows with experience in hospital administration and/or management to serve for 1 year as management mentors for the medical director and hospital management teams in the 14 hospitals. The Yale–Clinton Foundation Fellows were assigned in teams, with a full-time Post-graduate Fellow assigned to each of the 14 hospitals and Senior Fellows splitting their time between two hospitals. The Fellows worked on a daily basis in the hospitals as management mentors and lived for 12 months in the area in which the hospitals were located. During the year, the Ethiopian hospital management team members also participated in a professional development course offered by Yale Faculty on site. Course content included scientific problem-solving methods, problem identification and prioritization (e.g. use of flow charts, decision matrices and fishbone diagrams), project planning and evaluation, human resource management and leadership skills, working with governing boards and budgeting. At the end of the course, participants received the Yale Certificate in International Health Care Management. In collaboration, the Fellows, together with each hospital management team and in some cases with the help of RHB representatives, conducted a baseline needs assessment for the hospital, completed a problem identification and prioritization exercise and collectively identified an implementation plan and evaluation measures for completing at least one management project within their respective hospital.


We measured the change in the presence of 75 key management indicators from the time of baseline needs assessment (August 2006) to the follow-up 10 months later (May 2007). For each management indicator, the hospital reported it as ‘present’ (item = 1) or ‘absent’ (item = 0). Improvement was characterized as the hospitals having rated the indicators as ‘0’ at baseline and ‘1’ at follow-up. The key management indicator checklist was designed collaboratively by Yale University, the CHAI Fellows and the Federal Ministry of Health and, subsequently, formed the basis of the ‘Blueprint for Hospital Management in Ethiopia’ [17]. The checklist included a total of 75 indicators in the following domains: human resources, quality management, medical records and patient flow, nursing standards and practices, infection prevention and control and financial management.

In addition, we compared the Fellows' written reports of their perception of the skills and confidence of the medical directors and other hospital management staff at the 14 hospitals near baseline (September 2006) and follow-up (June 2007). Similarly, we compared the medical directors, and key managers' reports regarding the skills and contributions of the Fellows to the hospital environment and the overall program near baseline (September 2006) and follow-up (June 2007). Survey instruments employed a set of closed-ended Likert scale items and indicated perceptions of management skills and contributions. Instruments are available from the authors upon request.

Data analysis

We used descriptive statistics to summarize the percentage of hospitals with each management indicator present in August 2006 compared with May 2007. In addition, we summarized the surveys of Fellows and Ethiopian management teams using frequency analysis of the 4-point Likert scale responses and grouped responses into a binary variable (i.e. combining the two positive and two negative responses), with the data from baseline and from follow-up for each survey. We did not conduct statistical testing of changes in the prevalence of indicators as this was an exploratory study, and we lacked adequate statistical power given the limited sample size of the 14 hospitals. Qualitative comments recorded in the surveys were content analyzed and sorted by topic area.


Hospital sample

Half of the hospitals (n = 7/14) were located in the capital of Ethiopia, Addis Ababa. The remainder was split quite evenly between other urban (21.4%) or smaller city areas (28.6%). The mean number of beds was 240 per hospital, although the number ranged substantially from 74 beds in one hospital to >500 beds in another hospital. Overall, the average occupancy rate was 83%.

Hospital management indicators

We found improvement in 45 of the 75 (60%) key management indicators between August 2006 and May 2007 (Table 1). Improvements were apparent in multiple domains including human resources, medical records and patient flow, nursing standards and practice, infection prevention and control, quality management and financial management. The areas in which performance had been particularly low had the greatest improvement including conducting employee training, implementing nurse assessments upon admission and creating infection prevention manuals, committees and policies. Also, substantial increases were seen in the prevalence of quality improvement coordinators/teams, use of channels in the hospitals to report quality problems and employee participation in quality projects. The change in the percentage of hospitals from August 2006 to May 2007 that had the management indicator present reflected a total of 105 improvements across all 14 hospitals, which is a per-hospital mean of 7.5 (105/14) improvements, with SD equal to 5.9 improvements. Three hospitals (one in Addis and two in the rural areas) had at least 15 improvements, whereas four hospitals (two in Addis Ababa, one in the semi-urban areas and one in the rural areas) had fewer than five improvements.

View this table:
Table 1

Percentage of hospitals responding yes for the presence of each management indicator, for the 45 indicators that improved

% (Aug 2006)% (May 2007)
Human resources
 Hospital has a human resource department71.485.7
 Hospital performs employee training28.650.0
 Hospital uses performance evaluation system28.642.9
Medical records and patient flow
 Hospital has patient master index record numbers50.057.1
 Medical record contains
  Front sheet (e.g. patient name, contacts, etc.)64.385.7
  Doctor's order sheet78.692.9
  Progress notes for all disciplines57.171.4
  Laboratory results78.692.9
  Radiology reports57.171.4
  Consultation summaries50.064.2
  Discharge sheet71.478.6
 Patient flow has been mapped for
  Transport from home to hospital71.485.7
  Medication ordering, administering and documenting85.792.9
  Laboratory test ordering85.792.9
  Radiology test ordering64.378.6
  Referrals to and from other providers28.635.7
Nursing standards and practices
 Nurses perform the following
  Patient assessment upon admission21.450.0
  Draw blood samples71.485.7
  Document route, time and dose of medication78.685.7
  Teach patient/family about diagnosis and treatment64.471.4
 Hospital has standard for nurses' checks of patient status21.435.7
 Hospital has nurse-to-nurse reports at end of shifts64.371.4
 Nurses are licensed85.792.9
Infection prevention and control
 Hospital has infection prevention manual42.985.7
 Hospital has infection prevention committee28.664.3
 Hospital has infection prevention practitioner28.657.1
 Hospital has department-specific policies7.135.7
 Hospital follows standard (universal) precautions28.642.9
 Hospital uses cleaning process to minimize contamination21.435.7
 Clean and dirty items are stored separately57.178.6
 Hospital uses disinfectants57.178.6
 Hospital monitors hospital-acquired infections0.014.2
Quality management
 Hospital has mission statement85.792.9
 Hospital has continuing education for physicians42.957.1
 Hospital has implemented quality improvement programs35.771.4
 Hospital has quality improvement coordinator/team28.657.1
 Most employees are aware of quality improvement program14.321.4
 Hospital has set goals for quality28.635.7
 Most employees know hospital quality goals7.114.3
 Hospital has channels to report quality problems7.135.7
 Employees participate in quality projects14.371.4
 Hospital uses data feedback processes for quality21.428.6
Financial management
 Hospital has yearly budget plan92.9100.0
 Hospice has income statement78.692.9
 Hospital has statement of cash flow71.485.7

A total of 30 of the 75 (40.0%) management indicators surveyed did not increase or were reportedly lower in May 2007 (Table 2). In almost all of these areas in which no improvement was apparent, the baseline prevalence of the indicators was relatively higher (i.e. there was less opportunity for improvement than for the set of indicators where improvement was apparent).

View this table:
Table 2

Percentage of hospitals responding yes for the presence of each management indicator, for the 30 indicators where there was no improvement

% (Aug 2006)% (May 2007)
Human resources
 Hospital has organizational chart85.771.4
 Hospital has job descriptions85.785.7
 Hospital has compensation program describing salary/benefits57.150.0
Medical records and patient flow
 Medical record contains
  Medication administration record71.471.4
  Case management notes28.628.6
 Patient flow has been mapped for
  Patient admission/transfer78.657.1
  Medical record number assignment85.778.6
  Patient discharge78.678.6
  Patient billing85.778.6
Nursing standards and practices
 Nurses perform the following
  Insert intravenous catheters92.992.9
  Insert Foley catheters92.978.6
  Administer medications92.992.9
  Document clinical findings64.350.0
  Assist in ambulating/dressing/bathing patient71.471.4
  Document medical order carried out64.364.3
  Nurses report significant findings to physician92.985.7
  Nurses report to physician if medical order not effective78.664.3
Infection prevention and control
 Hospital has liquid soap available0.00.0
 Hospital has clean towels for drying hands7.17.1
 Hospital has alcohol rub available21.421.4
 Hospital has sinks with running water in all clinical areas57.150.0
Quality management
 Physicians are licensed100.092.9
 Hospital tracks quality or volume by department44.037.5
 Hospital has continuing education for nurses50.050.0
Financial management
 Hospital conducts budget variance analysis64.350.0
 Hospice tracks overall volume over time71.464.3
 Hospital record revenue sources for organization78.471.4
 Hospital has balance sheet71.471.4
 Hospital reports volume data to Ministry of Regional Bureau85.778.6
 Hospital reports financial data to Ministry of Regional Bureau92.985.7

The May 2007 data highlight key areas in which performance remains quite low, with fewer than 20% of hospitals having the relevant management indicators in place. These include employee knowledge of the quality goals of the hospital, surveillance on hospital-acquired infections and availability of liquid soap and clean towels in patient-care areas.

Fellows' perceptions of changes in management skills of hospital leaders

We found that the management skills of the medical directors as perceived by the Yale–Clinton Foundation Fellows improved from August 2006 to May 2007 in several management domains, although their level of confidence in their management skills did not increase generally (Table 3). Importantly, the relationships between medical directors and several key constituent groups (i.e. RHB and hospital administrative staff) were perceived to have improved for several hospitals.

View this table:
Table 3

Percentagea of hospital leaders' skills rated by mentors as excellent or good by management area

% (Aug 2006)% (May 2007)
Skills are excellent/good in
 Human resources37.541.7
 Hospital operations44.453.8
 Nursing management55.646.2
 Infection prevention and control22.238.5
 Quality improvement33.358.3
 Budgeting and finance14.345.5
Leaders' confidence level is excellent/good in
 Human resources60.061.5
 Hospital operations60.061.5
 Nursing management70.061.5
 Infection prevention and control60.046.2
 Quality improvement50.061.5
 Budgeting and finance33.361.5
Leaders' relationship is excellent/good with
 Regional health bureau25.066.7
 Hospital clinical staff70.061.5
 Hospital administrative staff50.069.2
  • aCalculated from number of surveys completed, which varied by item from 10–14.

Ethiopian hospital leaders' perceptions of Fellows' mentorship skills

In general, the Ethiopian hospital leaders, which included medical directors, matrons and administrators, perceived the Fellows' mentorship and interpersonal skills (e.g. listening, teaching and coaching) positively. These generally positive perceptions did not change appreciably from the beginning to the end of the program (Table 4). In data not shown, >90% of the hospital leaders reported themselves to be ‘very satisfied’ with the quality improvement projects in their hospitals; additionally, >90% of these leaders reported feeling ‘excellent’ or ‘good’ about the Hospital Management Initiative and believed the program would positively benefit other Ethiopian government hospitals.

View this table:
Table 4

Percentagea of mentors' skills rated by hospital leaders as excellent or good

Mentoring skill area% (Aug 2006)% (May 2007)
Listening skills93.0100.0
Teaching and coaching skills90.094.7
Ability to understand local context91.189.5
Establishing trust93.394.7
Showing courtesy and respect95.592.1
Interpersonal skills93.297.3
  • aCalculated from number of surveys completed, which varied by item from 10–14.


We found marked changes in several management indicators over 10 months of intervention using a partnership–mentoring model, EHMI, in Ethiopian hospitals. About 60% of the management indicators surveyed showed some improvement in the domains of human resources, medical records, nursing standards and practice, infection prevention and control, quality management and financial management. Although improvements were not uniform across all hospitals and across all indicators, the EHMI training and mentoring in quality improvement seemed to have a positive overall effect in key areas. This finding is consistent with other research showing that quality improvement efforts can be effective in lower-resource countries [7, 1015], although such work has not been previously conducted in Ethiopia, and in such a wide geography as six regions and two city administrations of the country.

An important aspect of successful healthcare management is the ability of leaders to work collaboratively and effectively with other groups in the system. In the hospital setting, such groups include the RHB or city health bureau and the hospital clinical and administrative staff. Our data suggest that the hospital leaders participating in this intervention were able to improve their relationships with these key constituent groups. These relationships may foster longer-term improvements in the management and integration of the hospital with its regional or city healthcare system.

Several aspects of this program are noteworthy. First, the program was conceived of, established by and supported by the Minister of Health, who attended training sessions and met regularly with the EHMI team to support their work. This support from the governmental and regulating body for hospitals was critical to facilitating changes. The Ministry provided substantial leadership in retaining focus on hospitals' efforts, motivating RHB to engage in hospital improvement activities and to create supporting regulation to enable hospitals to facilitate the recommended management actions. Without such in-country engagement, quality improvement efforts alone are less likely to be effective in strengthening health systems.

Second, the program provided more than technical support for the hospital leaders; it provided day-to-day mentorship with a committed international team of mentors who were selected to coach hospital leaders in management and quality improvement skills. In large part, this is because management is a process that is learned over time through trial and error, rather than a specific technical expertise that can be taught didactically. The management skills included learning to define a problem, set priorities, identify root causes, delegate to others, develop common goals, utilize flow chart processes and provide corrective feedback as needed. The success of this type of coaching requires strong and trusting relationships that are best developed over time and through regular face-to-face contact. Thus, the intervention demonstrates that the partnership–mentoring model can be an effective approach to management improvement efforts.

Third, the program reflected a partnership in its own governance, with strong involvement of a non-governmental organization and a university. Together, these partners provided the experience and expertise as well as the academic perspective to extract general lessons from specific projects in order to replicate these in other hospitals and contexts. For instance, the successful quality improvement projects contributed to a Blueprint for Hospital Management in Ethiopia [17], which characterizes eight key domains of ‘best practice’ in hospital management in resource-limited settings. This resource allows the lessons learned in a single quality improvement project to be translated into projects in other settings where senior hospital and governmental leaders foster replication. In addition, the partnership has led to the development of a new Master of Hospital Administration degree through a partnership with Jimma University and Yale University, as a way to professionalize the role of hospital management in the country. Such an educational track in the country can lead to sustained improvement in management, as a new cadre of hospital leaders are trained to strengthen this critical aspect of the healthcare system.

Our findings reflect the first study known to us that documents a countrywide effort in quality improvement focused on hospital management in Africa. Although the data suggest important improvements, they should be interpreted in the light of several limitations. First, the sample sizes for all analyses were limited. The study is therefore exploratory and descriptive without adequate statistical power to test the statistical significance of changes over time or between regions or hospitals. In addition, ratings of medical directors' and Fellows' skills were based on respondent perceptions. Nevertheless, the study does provide some initial information in an area that has not been previously evaluated and, therefore, is useful as an alternative model to help health systems gain greater attention. Second, turnover among hospital leaders during the year was unavoidable. As a result, the changes in hospital leaders' management skills reflect two cross-sections of leaders rather than a single cohort over time. Nevertheless, the quality improvement efforts resulted in positive changes despite such management turnover, which can slow processes of quality improvement. Third, it might be possible that data were misrepresented by hospitals in their evaluations; however, these evaluations were conducted collectively by the Yale–Clinton Foundation Fellows, the management teams, and in some cases, RHB representatives. Therefore, it would be difficult to misrepresent data markedly. Fourth, the study was conducted in Ethiopia, and results may differ in other countries. However, we did include hospitals from a range of urban and less urban settings suggesting that the approach can be effective in diverse settings. Last, while we experienced improvements in the first year of the program, sustained improvements will require additional strategies in future years of the program.

Sustaining the improvements in hospital management and further scaling-up the efforts are critical goals in the future years of the EHMI. Therefore in the second and third years of EHMI, additional cohorts of Ethiopian hospital leaders, who are appointed as Chief Executive Officers of the hospital, are enrolled in a Master of Hospital and Healthcare Administration (MHA) degree program, which requires for graduation a masters' thesis on quality improvement efforts in the hospital. At the same time, EHMI will seek greater alignment of RHB and FMOH goals regarding hospital improvement. To foster this, the partnership–mentoring model will be applied in the next 2 years at the RHB level, as it was applied at the hospital level during the first year of the program. Although external technical support to hospitals will still be needed, building leadership capacity in-country and government-hospital alignment are central to scaling up the improvements experienced in the first year of EHMI.

Our study, with existing evidence from other countries, contributes to growing research literature indicating that quality improvement efforts can be effective in making positive management changes in developing regions of the world. Nevertheless, experience even in higher-income countries in Europe [1820] and in the USA [2123] indicates that progress in quality improvement can be slow and its success depends on the commitment of senior organizational leaders to sustain improvement efforts. Although the program is implementing strategies to foster such senior leadership, longer term follow-up is needed to evaluate the replication and sustaining of these early accomplishments.


Patrick and Catherine Weldon Donaghue Medical Research Foundation Donaghue Investigator Award and the Clinton HIV/AIDS Initiative.


The authors are grateful to the committed efforts of the 2006–07 Yale-Clinton Foundation Fellows: Jennet Arcara, Elizabeth Arend, Charles Borden, Jaweer Brown, Lisa Browne, Rebecca Zewdie, Saran Ellis, Salem Fisseha, David Gervilla, David Hartley, Temi Ifafore, Mike Ikpe, Chris Kenyhercz, Mariama Samba-Koroma, Mary Anne Lally-Graves, Turkan Nabi, Steven Neri, Jackie Omotalade, Corrie Paeglow, David Ranney, Wendy Schwartz, Tara Thomas, Andre Villanueva, Zeleka Yeraswork and the CHAI team in Ethiopia, and Peter Huffman, who provided support for this project. The paper has benefited from thorough editing by Avital Havusha.


View Abstract