Mehtsun WT, Weatherspoon K, McElrath L, Chima A, Torsu VEK, Obeng ENB, Papandria DJ, Mehes MM, Ortega G, Hesse AAJ, Sory E, Perry H, Sampson J, Anderson J, Abdullah F. Assessing the Surgical and Obstetrics-Gynecology Workload of Medical OfficersFindings From 10 District Hospitals in Ghana. Arch Surg. 2012;147(6):542-548. doi:10.1001/archsurg.2012.449
Author Affiliations: The Johns Hopkins Bloomberg School of Public Health (Mss Mehtsun and McElrath, and Drs Weatherspoon, Chima, Perry, and Abdullah), and Departments of Surgery (Drs Papandria, Ortega, Perry, and Abdullah and Ms Mehes), Anesthesia (Dr Sampson), and Obstetrics and Gynecology (Dr Anderson), The Johns Hopkins University School of Medicine, Baltimore, Maryland; and Department of Surgery, Korle Bu Teaching Hospital (Dr Hesse), and Ghana Health Service (Dr Sory), Accra, Ghana. Mr Torsu and Ms Obeng are students at the University of Ghana Medical School, Accra.
Hypothesis Surgical and obstetrics-gynecology (Ob-Gyn) workload of medical officers (MOs) is substantial and may inform policies for training investment and surveillance to strengthen surgical care at district hospitals in Ghana.
Design Observational study.
Setting Academic research.
Participants Using standardized criteria, 12 trained on-site observers assessed the surgical and Ob-Gyn workload of MOs at 10 district hospitals in each of 10 administrative regions in Ghana, West Africa. The number of patients seen by MOs and the time spent managing each patient were recorded. According to each patient's diagnosis, the encounters were categorized as medical/nonsurgical, Ob-Gyn, or surgical.
Main Outcome Measures The proportions of patients having Ob-Gyn and surgical conditions and the time expended providing care to Ob-Gyn and surgical patients.
Results Of the observed patient encounters, 1600 (64.5%) were classified as medical or nonsurgical, 514 (20.7%) as Ob-Gyn, and 368 (14.8%) as surgical (9.0% nontrauma and 5.8% trauma). The most common diagnosis among Ob-Gyn patients was obstetric complication requiring cesarean section. The most common diagnosis among surgical patients was inguinal hernia. Medical officers devoted 24.8% of their time to managing Ob-Gyn patients and 18.9% to managing surgical patients (which included 5.4% for the management of traumatic injuries).
Conclusions Surgical and Ob-Gyn patients represent a substantial proportion of the workload among MOs at district hospitals in Ghana. Strategies to increase surgical capacity at these facilities must include equipping MOs with the appropriate training and resources to address the significant surgical and Ob-Gyn workload they face.
Surgical conditions constitute more than 7% of the disease burden in sub-Saharan Africa, yet surgical care has been neglected on the public health agenda in Africa and other low- and middle-income regions.1,2 Globally, an estimated 234 million major surgical procedures are performed annually, yet the poorest one-third of the world's population benefit from only 3% to 5% of these interventions.3 The available literature documenting a lack of access to emergency and essential surgical care in sub-Saharan Africa illustrates a need for surgical care evaluation through a public health framework.1,3- 5
Surgical care is increasingly recognized as critical to the prevention and treatment of disease. Male circumcision is a major public health intervention: it significantly reduces the risk of human immunodeficiency virus infection among men in sub-Saharan Africa, particularly among at-risk individuals.6 Recent evidence shows that the provision of surgical care can be cost-effective in low- and middle-income countries, discounting views that it is too expensive and is excessively resource intensive.7,8 Building surgical capacity at district hospitals has been championed as the most effective means to improve access to emergency and essential surgical care because these hospitals are usually the first point of access to surgical care in sub-Saharan Africa and in other low- and middle-income countries.5,9,10 Attaining suitable approaches to increase surgical and obstetrics-gynecology (Ob-Gyn) capacity necessitates a better understanding of the surgical and Ob-Gyn workload at the district level.
Ghana, West Africa, was chosen as the location for this study and represents an ideal locale because of existing collaborative relationships between the study team and physicians in Ghana and the Ghana Health Service. Furthermore, Ghana's National Healthcare Insurance program and widespread district hospital system are an ideal setting for research on surgical care at the district level. District hospitals provide medical care under the supervision of a medical officer (MO). Medical officers are physicians who have completed at least a 2-year internship (or “housemanship”) following graduation from medical school. The housemanship program consists of 6-month rotations in internal medicine, pediatrics, surgery, and Ob-Gyn. The surgical rotation is spent in different surgical subspecialty units.11
In a study assessing surgical and Ob-Gyn capacity of district hospitals in Ghana, Choo et al12 surveyed infrastructure characteristics, personnel, equipment, and surgical, obstetric, and anesthesia procedures at 17 hospitals in Ghana. They found that more than 75% of hospitals had adequate resources to provide general patient care and basic intraoperative care; however, only 29% had a formally trained surgeon, and only 36% had a formally trained obstetrician available. Also, 13 of 14 MOs interviewed had no formal surgical training beyond housemanship, during which they had performed only 4 to 8 major surgical procedures.11
The present observational study assessed the workload of MOs at 10 district hospitals in Ghana (1 in each of 10 administrative regions). Medical officers were observed across the spectrum of their clinical duties, including inpatient ward rounds, outpatient clinic, operative cases, and emergency department consultations. The objectives of the study were to determine the proportion of patients seen by MOs with surgical and Ob-Gyn conditions and the time expended providing care to these patients.
The study was conducted in Ghana, a low-income West African country with an annual per capita gross domestic product of $1098 and a population of 23.8 million.13 The administration of Ghana is divided into 10 regions and further into 110 districts. Government health services are administered by the Ghana Health Service, which operates at national, regional, and district levels.14 Community health centers, the smallest unit of the health delivery system, provide basic preventive and curative services. District hospitals provide preventive and curative services (including surgical care), serve populations of 100 000 to 200 000, and are the first referral center from community health centers.14 Regional hospitals serve as referral centers for district hospitals and provide more specialized health care.
Ten district hospitals, 1 from each of 10 administrative regions (Apam, Begoro, Ashanti, Brong Ahafo, Northern, Dodowa, Volta, Upper East, Western, and Upper West regions), were selected for inclusion in the study. The district hospitals represent a convenience sample based on existing relationships with physicians in Ghana and the availability of MOs to participate in the study. Although the facilities assessed do not constitute a representative sample of Ghanaian district hospitals, they serve large patient populations and are situated in diverse geographic locations in Ghana.
This study was independently approved by the Institutional Review Board of The Johns Hopkins Bloomberg School of Public Health and by the ethics review committee of the Ghana Health Service. The study team consisted of students from The Johns Hopkins Bloomberg School of Public Health (W.T.M., K.W., L.M., and A.C.), each of whom worked in tandem with a Ghanaian team member recruited from the University of Ghana Medical School (V.E.K.T. and E.N.O.O.). Twelve team members were trained in the same standardized observation technique. Time and patient assessment methods were pilot tested by the entire observation team for 2 days before study initiation to standardize patient encounter timing procedures and patient categorization using the assessment tool. Interrater responses were compared, and all discrepancies and inconsistencies were reconciled during this time to ensure reliable results across facilities.
Data collection was conducted during a 14-day study period. On the basis of facility size, 1 or 2 study team members were placed in each district hospital. At the individual facilities, each study team member was assigned to observe 1 MO daily using the assessment tool during a period averaging 5 days (range, 4-6 days), including at least 1 weekend day. A 6-day observation period was achieved in 70.0% of hospitals during the study. This approach enabled simultaneous assessments at multiple district hospitals during the same period. At facilities where MOs outnumbered the study team, a different MO was randomly selected for observation each day of the observation period. A typical workday consisted of 8 to 12 hours, and no direct observation occurred overnight; however, an overnight log was kept to record the patients seen. Time expended was not recorded for these patient encounters.
At each facility, a standardized assessment tool was used to quantify the total time and the number and category of patients seen by MOs during the 14 days of data collection. The standardized data collection forms and observation techniques were developed by the study team before commencing the study. These were designed to collect facility-level data at each hospital (the numbers of beds, medical staff, and population served), patient demographic information (sex and age), and clinical details (diagnosis, referral decision, and diagnosis category). The patient diagnosis categories included the following 3 domains: Ob-Gyn, surgical (nontrauma and trauma), and medical or nonsurgical (including pediatric nonsurgical conditions). A list of conditions and disease states commonly encountered at district hospitals was compiled a priori across categories. Using guidelines by Bickler et al,15 surgical conditions were defined as those “ideally” requiring the expertise of a surgically trained physician. A similar definition was used to qualify Ob-Gyn conditions. Before the start of the study, the definitions and possible diagnoses for each category were reviewed and adjudicated by the study team and principal investigators (W.T.M., K.W., L.M., A.C., H.P., and F.A.). Patient diagnoses that did not fall under the surgical category or Ob-Gyn category and that did not require any incision or manipulation (eg, the setting of a fracture) were placed in the medical or nonsurgical category.
In cases in which multiple diagnoses were observed, patients were systematically grouped according to the first diagnosis listed by the MO. The length of the patient encounter was measured in minutes and seconds using a standardized technique with stopwatches. The start of the patient encounter was defined as the instant the MO came into contact with the patient or with related clinical information (including medical records, imaging, and others). The stopwatch was started at the initiation of the patient encounter, was paused during any interruption of the encounter, and was restarted when the encounter resumed. Finally, when the MO officially ended the encounter or moved on to the next patient, the stopwatch was stopped, and the time was recorded. Nonclinical duties were not assessed, and the encounters conducted in the absence of the study observers were excluded from measurement or analysis. Additional information collected included the following: hospital features, district population characteristics, distance to the nearest referral center, and the numbers of operating rooms, MOs and ancillary staff, and surgical specialists available at each hospital and total district hospitals in the region. Surgical specialists were defined as physicians who had completed residency training in general surgery or Ob-Gyn, including their respective subspecialties.
In assessing the MO workload, the following primary outcomes were assessed: the total number of patients in each defined diagnostic category, the total time dedicated to the management of patients in each category, and the total number of patients referred to other facilities for further treatment. The secondary outcomes were computed relative to the total number of patients and the total encounter time recorded and included the proportions of patients and time spent in each respective diagnostic category, as well as the proportion of patients referred. These values were calculated for individual district hospitals and in aggregate for all facilities assessed.
Descriptive analyses were performed. Summary statistics of the primary and secondary outcomes were reported. A sensitivity analysis was conducted to compare categorical proportions between all the patient encounters and those directly observed. All the analyses were performed using commercially available software (STATA, version 11; StataCorp LP).
The district population served by each hospital in the study ranged from 53 558 to 223 688 (Table 1). The number of district hospitals located within each region included in the study ranged from 4 to 27. The distance from district hospitals in the study to the nearest regional referral center ranged from 25 to 150 km. The number of MOs assigned to each district hospital ranged from 1 to 7, and at least 1 surgical specialist was on staff at 3 of the locations studied (Ob-Gyn in all 4 instances). In addition, the number of ancillary staff (including nurses, midwives, and student nurses) ranged from 26 to 140.
In total, data were collected on 2847 patient encounters at 10 district hospitals during the study period. Among all patients seen by MOs, 1734 (60.9%) were female, 800 (28.1%) were designated as pediatric (<18 years), and 1337 (47.0%) were outpatients. In total, 2482 (87.2%) of the recorded encounters were directly observed and timed by the study observers. Of the directly observed adult and pediatric patient encounters, 1600 (64.5%) were classified as medical or nonsurgical, 514 (20.7%) as Ob-Gyn, and 368 (14.8%) as surgical (9.0% nontrauma and 5.8% trauma) (Table 2). An analysis of diagnostic categorical proportions between the total patients encountered and those directly observed showed no statistically significant differences.
Among the directly observed encounters by MOs, 54% of their time was spent on inpatient ward rounds, 46% at outpatient clinics, 15% in the operating room, and 7% with patients in the emergency department. On average, 56.3% of the clinical encounter time of MOs was expended on medical or nonsurgical patients, 24.8% on Ob-Gyn patients, and 18.9% on surgical patients (13.5% nontrauma and 5.4% trauma) (Table 3). Operative time accounted for 35% of the total time spent with Ob-Gyn patients and for 36% of the total time spent with surgical patients. District hospitals with 100 beds or more had significantly more surgical encounters (P = .002), and MOs there spent a mean of 139.3 seconds less on surgical encounters than hospitals with fewer than 100 beds (P = .04). District hospitals without a trained surgeon had fewer surgical encounters than hospitals with at least 1 trained surgeon (P < .001). However, no statistically significant difference in the mean time spent on surgical encounters was observed between hospitals with vs those without a surgically trained MO.
Table 4 lists common diagnoses among all the patient encounters by diagnostic category. The most common diagnoses (by diagnostic category) were malaria (medical or nonsurgical), obstetric complication requiring cesarean section (Ob-Gyn), inguinal hernia (surgical nontrauma), and motor vehicle crash (surgical trauma).
To estimate the workload of MOs at district hospitals in Ghana, this study investigated surgical and Ob-Gyn care by recording the number of patients seen and by measuring the time expended. A review of the surgical and public health literature revealed no previous investigations using such methods to estimate surgical and Ob-Gyn workload. Similar methods have been used to assess physician efficiency in developed countries.16 The timed observation method in the present study represents a novel instrument for estimating the clinical workload of physicians working at the district hospital level in low- and middle-income countries.
The results show that surgical and Ob-Gyn patients accounted for more than one-third of patients seen by MOs and for almost half of an MO's typical workday in the hospitals observed. Operative time dedicated to the care of surgical and Ob-Gyn patients represented more than one-third of the total time spent on patients within each respective diagnostic category. These findings suggest that the management of surgical and Ob-Gyn patients represents a substantial proportion of the clinical duties of MOs at these district hospitals.
Throughout the 10 district hospitals observed, the most common diagnosis among Ob-Gyn patients was obstetric complication requiring cesarean section. The most common diagnosis among surgical patients was inguinal hernia. In a retrospective review of operative logs at 8 district hospitals in Uganda, Tanzania, and Mozambique, Galukande et al10 similarly found cesarean sections and uterine evacuations to be the most common Ob-Gyn cases; hernia repair and wound care accounted for 65% of all general surgical procedures. Likewise, a 5-year operative log review by Damien et al17 found cesarean sections to be the most common surgical procedure (41%) performed in the Kintampo district hospital in Ghana. Cesarean section was also the most common major surgical procedure in an assessment of operative caseloads across 21 district hospitals in Malawi.18 Our data are consistent with these findings.
In our study, a referral rate to tertiary medical facilities of 2.0% was observed among the patients seen. Several Ghanaian studies19- 21 have attributed the absence of a functional referral system to inadequate roads, poor ambulance services, and irregular transportation, accompanied by high default rates. This may explain the low referral rate found in the present study. Additional research is required to clarify this issue.
Previous investigations estimating global surgical disease burden and volume have used indirect measures and modeling. Debas et al1 applied estimates derived from a survey of 32 surgeons globally to the disability-adjusted life-year estimates of disease burden provided by the 2002 World Health Report and the Global Burden of Disease study.22- 24 A surgical condition was defined as “any condition that requires suture, incision, excision, manipulation, or other invasive procedure that usually, but not always, requires local, regional, or general anesthesia.”25 Weiser et al26 estimated and highlighted the global disparity of surgical care by modeling estimates based on surgical data from 56 of 192 World Health Organization member states: 73.6% of surgical procedures occurred in the wealthiest countries, which have only one-third of the world's population, while only 3.5% of all surgical procedures occurred among the poorest one-third. These pioneering studies have provided the first crude global estimates of surgical disease burden and access to surgical services.
Bickler and Spiegel27 underscored a need for the development of data-driven strategies to improve surgical services in low- and middle-income countries. The present study may serve as an approximation of the workload resulting from surgical and Ob-Gyn conditions relative to other prevalent diseases at district hospitals. Our definition of a surgical condition expands beyond that used by the Disease Control Priorities Project.1 As highlighted by Ozgediz et al,28 some common surgical conditions may not require an incision and in some instances may not need operative management, despite the need for a qualified surgical consultant; common conditions, such as acute abdominal emergencies and surgical infections, were excluded from the initial Global Burden of Disease study. Our findings demonstrate that among surgical and Ob-Gyn patient encounters, more than two-thirds of patients did not undergo an operative intervention at the time of observation. Our data support the call for a broader definition of surgical conditions to include all the aspects of patient care, not just the surgical procedure itself, for a more accurate understanding of conditions requiring surgical management.
Interpretations of these data reported herein should consider limitations that result from the methods used. Many patients in Ghana fail to seek care at health facilities because of economic constraints, inadequate transportation, and fear of or lack of confidence in the quality of care provided. These data herein reflect the characteristics of patients capable of accessing health care at a district hospital. Inherent variability also exists in MO workloads across hospitals owing to differences in case mix, patient populations, and facility characteristics (operating room capacity, diagnostic and surgical equipment availability, and the numbers of physicians, specialists, and hospital beds, and others). Seasonal variations in disease conditions may have resulted in an underestimation or overestimation of disease categories because our assessments were conducted solely from January 4 through 19, 2011.26 Patients with multiple diagnoses were assigned to clinical categories based on the order of documentation, which may have biased the classification of patients with complex disease into certain categories. Finally, inferences related to quality of care, postoperative complications, or postoperative mortality are beyond the scope of this analysis.
Among the district hospitals studied, only 4 physicians had received formal surgical training (Table 1). This contrasts with the high concentrations of surgical specialists in regional and tertiary health facilities in Ghana.17 Most of Ghana's population are located in rural areas, with limited access to these specialists, further underscoring the need to allocate adequate resources and training interventions at the district level.15,26,29 Health policy makers must be guided by regional and district data to establish adequate and appropriate interventions.27,30
In conclusion, this study demonstrates a novel approach to assessing the workload of MOs by addressing not only patient mix but also the time allotted to patient care. Our findings may inform policies for training, investment, and surveillance to strengthen surgical care at district hospitals in Ghana. As initiatives to enhance surgical capacity in low- and middle-income countries emerge, the need for evidence-driven solutions to existing regional disease burden is paramount. Surgical care is a public health priority. Expanding the provision of surgical services in low- and middle-income countries will require the identification of strategies that address health system weaknesses and ensure adequate infrastructure and efficient use of human resources.
Correspondence: Fizan Abdullah, MD, PhD, Department of Surgery, The Johns Hopkins University School of Medicine, 600 N Wolfe St, Harvey Room 319, Baltimore, MD 21287 (firstname.lastname@example.org).
Accepted for Publication: December 14, 2011.
Author Contributions:Study concept and design: Mehtsun, Weatherspoon, McElrath, Chima, Papandria, Ortega, Hesse, Sory, Perry, Sampson, Anderson, and Abdullah. Acquisition of data: Mehtsun, Weatherspoon, McElrath, Chima, Torsu, Obeng, and Hesse. Analysis and interpretation of data: Mehtsun, Weatherspoon, McElrath, Chima, Papandria, Mehes, Ortega, Perry, Sampson, and Abdullah. Drafting of the manuscript: Mehtsun, Weatherspoon, McElrath, Chima, Torsu, Obeng, Papandria, and Abdullah. Critical revision of the manuscript for important intellectual content: Mehtsun, Weatherspoon, McElrath, Chima, Papandria, Mehes, Ortega, Hesse, Sory, Perry, Sampson, Anderson, and Abdullah. Statistical analysis: Mehtsun, Weatherspoon, McElrath, Chima, Papandria, and Ortega. Obtained funding: Mehtsun, Weatherspoon, McElrath, Chima, and Sampson. Administrative, technical, and material support: Mehtsun, McElrath, Torsu, Obeng, Papandria, Mehes, Ortega, Hesse, Sory, and Abdullah. Study supervision: Mehtsun, Chima, Torsu, Perry, Sampson, and Abdullah.
Financial Disclosure: None reported.
Funding/Support: This study was supported by Humanity First International and by the generous contributions of family and friends.
Additional Contributions: Logistical support was provided by Simon K. Antey (retired administrator officer at the World Health Organization) and family. The following students from the University of Ghana Medical School assisted with data collection: Jonathan Wadeyir Abesig, BSc; Francis Vinkpenuba Wuobar, BSc; Colin Kwakofi, BSc; Vincent Jessey Ganu, BSc; Roselyn Edzordzie Glover, BSc; Nonterah Engelbert Adamwaba, BSc; Ramsys Nii Odartei Mills, BSc; and Bright Kyei Wiredu, BSc.