District-level surgery in Uganda: Indications, interventions and perioperative mortality Jenny L€ ofgren, MD,a Daniel Kadobera, MSc(Med),b Birger C. Forsberg, MD,c Jude Mulowooza, MD,d Andreas Wladis, MD, PhD,e and P€ar Nordin, MD, PhD,f Ume a, Solna, and Stockholm, Sweden, and Iganga, Uganda
Background. The world’s poorest 2 billion people, benefit from no more than about 3.5% of the world’s operative procedures. The burden of surgical disease is greatest in Africa, where operations could save many lives. Previous facility-based studies have described operative procedure caseloads, but prospective studies investigating interventions, indications and perioperative mortality rates (POMR), are rare. Methods. A prospective, questionnaire-based collection of data on all major and minor operative procedures was undertaken at 2 hospitals in rural Uganda covering 4 and 3 months in 2011, respectively. Data included patient characteristics, indications for the interventions performed, and outcome after surgery. Results. We recorded 2,790 operative procedures on 2,701 patients. The rate of major operative procedures per 100,000 population per year was 225. Patients undergoing major operative procedures (n = 1,051) were mostly women (n = 923; 88%) because most interventions were performed owing to pregnancy-related complications (n = 747; 67%) or gynecologic conditions (n = 114; 10%). General operative interventions registered included herniorrhaphy (n = 103; 9%), exploratory laparotomy (n = 60; 5%), and appendectomy (n = 31; 3%). The POMR for major operative procedures was 1% (n = 14) and was greatest after exploratory laparotomy (13%; n = 8) and caesarean delivery (1%; n = 4). Most deaths (n = 16) were a result of sepsis (n = 10–11) or hemorrhage (n = 3–5). Conclusion. The volume of surgery was low relative to the size of the catchment population. The POMR was high. Exploratory laparotomy and caesarean section were identified as high-risk procedures. Increased availability of blood, improved perioperative monitoring, and early intervention could be part of a solution to reduce the POMR. (Surgery 2015;158:7-16.) From the Department of Surgery and Perioperative Sciences,a Ume a University Hospital, Ume a, Sweden; School of Public Health, Iganga/Mayuge Health and Demographic Surveillance Site,b Iganga, Uganda; Department of Public Health Sciences,c Karolinska Institute, Solna, Sweden; Iganga General Hospital,d € e S€o dersjukhuset, Karolinska Iganga, Uganda; Department of Clinical Science and Education (KI SOS), Institute, Stockholm, Sweden; and the Department of Surgery and Perioperative Sciences,f Ume a University Hospital, Ume a , Sweden
MORE THAN 2 BILLION PEOPLE, OR 35% OF THE WORLD’S POPULATION, live in countries where only 3.5% of the world’s operative interventions are carried This work was conducted as a collaboration between Makerere University, Uganda, and Ume a University Hospital and the Karolinska Institute, Sweden. Funding was received from the Swedish Society of Medicine as well as the Golje Foundation, Sweden. The authors of this manuscript have no conflicts of interest in relation to the present manuscript or its contents. Accepted for publication March 20, 2015. Reprint requests: Jenny L€ ofgren, MD, Department of Surgery and Perioperative Sciences, Ume a University Hospital, 901 85 Ume a, Sweden. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.03.022
out.1 There is a large unmet need for surgery, and many lives could be saved by increasing the number of operative interventions.2-5 Africa carries the greatest burden of surgical disease in the world.6 Traditionally, surgery has been considered to be too costly and ineffective from a public health perspective, but cost and cost-effectiveness analyses disprove this viewpoint.7,8 Operative intervention has been identified as a neglected low-cost opportunity in both South Asia and sub-Saharan Africa, and attention to this aspect of health care in low- and middle-income countries is increasing.9-12 Reliable data on baseline surgical production are required if one is to plan for and evaluate the results of efforts to improve surgical services in lowresource settings. Comparing data on surgical output with data on the need for surgery in a SURGERY 7
8 L€o fgren et al
reference population is 1 way to measure the surgical service provision gap. Assessing postoperative outcomes will help to identify potential areas for quality improvement. One such outcome, the perioperative mortality rate (POMR), has been suggested recently as a possible indicator of accessibility to safe surgery and anesthesia.13 A number of retrospective, facility-based studies describing the caseload of surgical patients and operative procedures have been carried out in various locations in sub-Saharan Africa.14-17 In contrast, prospective studies that also include indications for operation and its outcome are scarce. The aim of the current study was to provide detailed information on operative procedures, their indications, and outcomes at 2 district hospitals in rural Uganda. METHODS Location. Data collection was carried out prospectively at the Iganga General Hospital (IGH) and the Buluba Mission Hospital (BMH) in Eastern Uganda in 2011. These hospitals were selected because they provide surgery services for the population of a so-called Health and Demographic Surveillance Site, where a study on groin hernia prevalence was carried out at the same time as the present study.4 These 2 hospitals were the primary providers of major operative procedures for a catchment population of 1.5 million.18 More specialized operative and obstetric care is provided at the nearest regional referral hospital in Jinja, located 35 km by road from the town of Iganga. This distance and the associated costs pose a barrier to the generally poor population. Private facilities that are not registered or licensed to carry out major operations may do so anyway. This study presents information from the 2 hospitals that officially carry out major operations. BMH, a nonprofit organization, does not have a defined catchment area, but basically serves the same population as the IGH. The IGH had 150 beds, 1 major and 1 minor operating room, 11 medical doctors, and 85 nurses and midwifes at the time of the study. The BMH had 120 beds with 1 operating room, 4 medical doctors, and 30 nurses and midwifes. Anesthesia in both hospitals was provided by mid-level health care providers with 3 years of training in anesthesia. The 2 hospitals had 1 gynecologist/obstetrician each. The IGH had a vacant position for a surgeon, and the BMH had a visiting surgeon once a week. All district hospitals in Uganda have positions for 5 specialist medical doctors, one of which is a surgeon. Owing to financial constraints,
Surgery July 2015
however, most hospitals do not actually recruit these specialists. In this respect, the 2 hospitals are typical of this level of hospital in Uganda. Definitions. A major operative procedure was defined as incision, excision, manipulation, or suturing of tissue that normally requires regional or general anesthesia or profound sedation to control pain.19 All other operative procedures, suitable to be undertaken under local anesthesia or nerve block, were defined as minor. Owing to overlap in treatment and anesthetic practice, all orthopedic procedures were grouped as ‘‘minor procedures.’’ The most invasive orthopedic operation in this study was insertion of a pin for traction. POMR is expressed as a percentage, and was defined as in-hospital mortality owing to any cause during surgery or within 30 days after an operative procedure divided by the number of individuals having undergone an operation. Data collection. Data were collected prospectively in 2011 using 2 pretested questionnaires, one for data collection in the operating room and one for data collection on the ward. Staff already employed at the 2 hospitals were trained to carry out data collection. Data on all major operative procedures undertaken at the 2 hospitals were collected over 4 months, and data on all minor procedures were collected simultaneously, but for 3 months only because they were much greater in number. The data collected were basic patient characteristics, type of procedure performed, indication, whether the patient was admitted or treated as an outpatient, and whether the patient was discharged, referred, or died. Complications were included, and there was also space to fill any further comments or information. J.L. and J.M. carried out quality control of the data collected. Forms filled in the operation room were crosschecked with the operating room registers and the forms filled in on the ward. If required, clinical notes from the wards and registers at the time of admission were checked to correct any inconsistency. Additional data on service provision at the 2 hospitals were retrieved from the hospitals’ annual reports. Ethical considerations. Ethical approval for the study was granted by the Institutional Review Board at School of Biomedical Sciences, Makerere University, as well as the Uganda National Council for Science and Technology. Approval for data collection was granted by the medical superintendent at each hospital. Data analysis. Double entry of the data was practiced to minimize errors. Basic statistical
L€o fgren et al 9
Surgery Volume 158, Number 1
calculations were performed using Excel 2013 (Microsoft Corporation, Redmond, WA).
RESULTS A total of 2,790 operative interventions on 2,701 patients were recorded. Of these, 1,051 patients underwent a major operative procedure. This corresponds with an annual rate of major surgery of 225 procedures per 100,000 in the catchment area population. Most patients were operated at the IGH (major, 749 [71%]; minor, 1,558 [94%]). Patients undergoing major operative procedures were mostly women (n = 913 [87%]) whereas the majority of patients undergoing minor operative procedures were men (n = 1,293 [78%]). Patients who had major operative procedures were older (mean age, 28.5 years) than those undergoing minor operative procedures (mean age, 18.4 years; P < .001). Almost all major procedures were carried out by general medical doctors. Only 17% (n = 180) were operated by either a qualified surgeon or gynecologist/obstetrician. Most minor procedures were performed by nurses, midwifes, and clinical officers. Clinical officers are midlevel providers with 3 years of clinical training. Table I shows the characteristics of the patients and the procedures carried out. Table II shows the numbers and proportions of major and minor operative procedures recorded. Table III shows the indications for the most commonly performed major operative procedures. Male circumcision, the most commonly performed minor operative procedure, was done for primary prevention of human immunodeficiency virus (HIV) in >99% of cases. The indications for the remaining minor procedures were standard. The most invasive orthopedic procedure was insertion of a traction pin in the tibial tuberosity in femur fractures. The in-hospital POMR was 0.6% (n = 16). Deaths occurred in both hospitals (13 in IGH and 3 in BMH). As shown in Table IV, death was most commonly seen after exploratory laparotomy (n = 8 [13%]) and caesarean delivery (n = 4 [0.8%]). Each death is described in detail in Table IV. Of 510 babies delivered via the 496 caesarean deliveries, 59 (12%) were either stillborn or died before leaving hospital. Hospital data revealed that overall maternal mortality in the 2 hospitals between July 2010 and June 2011 was 43 of 4,808 deliveries (0.9%) at the IGH and 10 of 1,042 deliveries (1%) at the BMH. Of the patients who died of sepsis after exploratory laparotomy, all but one had been prescribed
Table I. Characteristics of patients undergoing major and minor operative procedures at Iganga General Hospital and Buluba Mission Hospital
Characteristic
Major operative Minor operative procedures procedures (n = 1,051)* (n = 1,650)y
Female sex, n (%) 914 Age Mean, y (SD) 28.5 Children (US$1,000 per capita, confirming the grossly uneven global distribution of access to surgery.1,17 The annualized caesarean delivery rate of 99.2 per 100,000 persons per year suggests that services have increased since the 1980s when rates were reported to be around 25 per 100,000 per year in rural settings in Eastern Africa. The crude rate of
10 L€o fgren et al
Surgery July 2015
Table II. Distribution and rates of major operative procedures and distribution of minor operative procedures at Iganga General Hospital and Buluba Mission Hospital* Major surgery (n = 1,124)y Intervention Caesarean delivery Evacuation of uterus Herniorrhaphy (all kinds of abdominal wall hernia) Explorative laparotomy Hysterectomy Salpingo/oophorectomy or removal of tubal/ovarian mass Appendectomy Tubal ligation Repair of cleft lip or palate* Laparotomy for ectopic pregnancy Other Total
n (%)
Minor surgery (n = 1,666)z Rate per 100,000 persons per year
496 (44.1) 244 (21.7) 103 (9.2)
99.2 48.8 20.6
60 (5.3) 54 (4.8)
12.0 10.8
36 (3.2)
7.2
31 24 21 10
(2.8) (2.1) (1.9) (0.9)
6.2 4.8 4.2 2.0
45 (4.0) 1,124 (100)
9.0 224.8
Intervention Circumcision Repair of cuts and tears Incision and drainage of abscess Reduction of fracture Excision of granuloma of the umbilical chord Excision of mass Incision of tongue tie Reduction of dislocated joint Biopsy Removal of foreign body/implant Other Total
n (%) 768 (46.1) 293 (17.6) 238 (14.3) 138 (8.3) 36 (2.2) 35 (2.1) 22 18 16 15
(1.3) (1.1) (1.0) (0.9)
87 (5.2) 1,666 (100)
*A surgical camp for cleft lip and palate was undertaken by a plastic surgeon during the time of the study. It occurs at regular intervals, usually twice per year, but cleft palate repair is not a routine intervention provided by the hospitals at other times. yData on major operative procedures were collected over 4 consecutive months. zData on major minor operative procedures were collected over 3 consecutive months. Minor procedures are also undertaken in lower level health care facilities; therefore, reliable rates could not be generated. The number of surgeries exceeds the number of patients as some patients underwent >1 intervention at the same time. Total is not exactly 100% owing to round off error.
herniorrhaphy (21 per 100,000 persons per year), on the other hand, is lower than the 25 per 100,000 persons reported in East Africa, but similar to findings from the IGH and BMH in 2006.14,15 Considering the high prevalence of groin hernia found in the study setting, the annual hernia repair rate should ideally exceed the 130– 290 groin hernia repairs per 100,000 persons per year seen in high-income countries.4,20,21 Male circumcision has been shown to reduce the risk for men to contract HIV through sexual contact.22 Large efforts have been invested in increasing male circumcision, resulting in it being the most commonly undertaken operative procedure in this study. Incision and drainage of abscess was the third most commonly performed minor procedure. This is a simple but potentially lifesaving intervention; abscesses and soft tissue infections are the seventh leading cause of acute surgical mortality.23 Suturing of cuts and lacerations and reduction of fractures were other common minor procedures. Although conclusions about the burden of disease cannot be drawn from facility-based data, these findings indicate an high incidence of trauma in the community. Injuries contribute
greatly to the global burden of disease.24 This burden could be eased considerably through an increase in basic surgical services, such as those provided in the study setting.25 Major operative procedures on trauma patients were very rare, and many people who sustain injuries to the skull, chest, abdomen, or pelvis probably die before reaching a hospital. With future improvements in prehospital care, it is probable that the need for major surgery resources in this group of patients will increase considerably. POMR and cause of death. Saving lives is a primary goal of any health care system. POMR and deaths audit are important tools for improving the quality of service delivery. Calculating POMR identifies procedures that carry the highest mortality rates. Identifying cause of mortality will help to set priorities for both capacity improvement and future research. Auditing of POMR and complications is an inexpensive way of evaluating the quality of surgical practice. In the present study, the POMR varied with the type of operative procedure undertaken. Eight deaths were recorded after emergency exploratory laparotomy, resulting in the highest POMR (13%) in this study. The commonest indication for
L€o fgren et al 11
Surgery Volume 158, Number 1
Table III. Indications for major operative procedures Intervention Caesarean delivery (n = 496)
Evacuation (n = 244)
Hernia repair (n = 103)
Explorative laparotomy (n = 60)
Hysterectomy (n = 54)
Salpingo/oophorectomy or removal of mass in salpinx or ovaries (n = 37)
Indication Prolonged or obstructed labor Previous caesarean delivery Cephalopelvic disproportion Malpresentation Fetal distress Antepartum hemorrhage Other Incomplete abortion Bleeding Retained placenta or membranes Other Groin hernia Other abdominal wall hernia Other Gastrointestinal obstruction Peritonitis Abdominal pain Unknown Other Myoma Bleeding Prolapse Cancer (suspected or known) Other Ovarian cyst In conjunction with hysterectomy Other
n (%) 196 156 143 66 43 31 146 210 161 25 9 85 12 2 20 20 10 4 10 21 12 10 9 11 19 9 9
(39.6) (31.5) (28.8) (13.3) (8.7) (6.3) (29.4) (86.1) (66.0) (10.2) (3.7) (82.5) (11.7) (1.9) (33.3) (33.3) (16.7) (6.7) (16.7) (38.9) (22.2) (18.5) (16.7) (20.4) (52.8) (25.0) (25.0)
Totals may exceed 100% as surgery could be performed owing to >1 indication.
operative intervention was peritonitis and/or gastrointestinal obstruction, the third leading cause of death after emergency general surgery in the world.23 Emergency laparotomy is also hazardous in the high-income setting, with an overall mortality rate of 14.9% as reported in a study from the UK in 2012.26 Another study stressed the importance of timely management, involvement of senior clinicians, accessibility of suitable radiology resources, and the availability of critical care in high-risk patients.27 The timing of operative procedures could probably be improved in the study setting; many patients present late. The other 3 factors mentioned may not be correctable for many years to come. Instead, improvements not requiring a large investment should be the first priority in settings with limited resources. The use of pulse oximetry, surgical checklists, and basic hygiene routines are examples of practices that can improve operative care and outcomes.28,29 The POMR after caesarean delivery was 0.8% (n = 4), which is slightly less than the overall maternal
mortality seen at the 2 hospitals. The 4 maternal deaths in this study were a result of hemorrhage or sepsis, both common causes of maternal death in East Africa.30 All 3 deaths owing to hemorrhage were treated inadequately in terms of blood transfusion. Whether or not these deaths could have been averted is unknown, but improving the availability of blood for such patients should help to reduce maternal mortality. Most deaths in the present study were a result of sepsis. Adequate antibiotic and fluid therapy should, therefore, play a central role in reducing the risk for death after operative procedures. Of the patients who died of sepsis, all but one had been prescribed intravenous fluids and antibiotics. We cannot say, however, whether or not this treatment was given in a timely manner, in adequate doses, and with evaluation of treatment effect. Stock outs are common and patients cannot always afford to buy the medicines required. This situation, together with low staffing levels, influences adherence to recommended treatment. In addition to the low-cost improvement measures
Intervention/ Patient nr. Caesarean delivery (n = 496) 1
Deaths, Age n POMR (y)
Indication for surgery/diagnosis after surgery
Surgical planning
Post operative complication
Cause of death (as stated in questionnaire)
Days between surgery and death
Comment
Most probable cause of death
—
—
—
—
—
—
—
—
1
Grand multipara
Sepsis
Hemorrhage
0
Hemorrhage
Hemorrhage
0
Also had perineal rupture which was repaired —
4
0.8
—
—
—
Emergency
—
2
—
—
Emergency
—
3
—
—
4
—
—
8
13.3
30 Obstructed labor and sepsis 20 Eclampsia and previous caesarean section 30 Two previous caesarean sections and in labor 23 Antepartum hemorrhage — —
—
—
2
—
—
3
—
—
4
—
—
5
—
—
6 7
— —
— —
Explorative laparotomy (n = 60) 1
Emergency Hemorrhage
Emergency
—
—
0
—
—
—
—
—
—
40 Peritonitis/fecal impaction 40 Peritonitis/ intestinal obstruction 48 Peritonitis/ peritonitis
Emergency
—
—
9
—
0
Emergency Deep infection
Sepsis
6
18 Large bowel obstruction 80 Small bowel obstruction
Emergency Deep infection
Sepsis
6
Emergency Deep infection
Sepsis
4
70 Sigmoid volvulus 70 Large bowel obstruction/ intestinal obstruction
Emergency Deep infection Emergency Wound dehiscence
Emergency Severe pain
— —
7 0
12 L€o fgren et al
Table IV. Perioperative mortality rates (POMR) per 100 interventions, characteristics, and cause of death in diseased patients
Hemorrhage
Hemorrhage —
Drained 4 L of pus Sepsis from abdomen Sepsis
Also underwent sigmoid resection
Sepsis
Sepsis Also had intraabdominal adhesions
Sepsis
Sepsis Sepsis
Surgery July 2015
(continued)
Intervention/ Patient nr. 8
Herniorrhaphy for groin hernia (n = 83) 1 Repair of cleft lip/ palate (n = 21) 1
Repair of cuts and tears (n = 293) 1
Wound debridement (n = 11) 1
Surgical planning
Post operative complication
Cause of death (as stated in questionnaire)
Days between surgery and death
Comment
40 Peritonitis and small bowel obstruction/ peritonitis and intestinal obstruction — —
Emergency
—
—
0
—
—
—
—
—
—
60 Obstructed groin hernia — —
Emergency
Hematoma
—
1
—
—
—
—
—
—
Elective
Aspiration
Aspiration
0
—
—
—
—
Emergency
—
—
0
—
—
—
—
Emergency
—
—
12
Deaths, Age n POMR (y)
Indication for surgery/diagnosis after surgery
—
—
1
1.2
—
—
1
4.8
—
—
Background. The world’s poorest 2 billion people, benefit from no more than about 3.5% of the world’s operative procedures. The burden of surgical disease is greatest in Africa, where operations could save many lives. Previous facility-based studies have described operative procedure caseloads, but prospective studies investigating interventions, indications and perioperative mortality rates (POMR), are rare. Methods. A prospective, questionnaire-based collection of data on all major and minor operative procedures was undertaken at 2 hospitals in rural Uganda covering 4 and 3 months in 2011, respectively. Data included patient characteristics, indications for the interventions performed, and outcome after surgery. Results. We recorded 2,790 operative procedures on 2,701 patients. The rate of major operative procedures per 100,000 population per year was 225. Patients undergoing major operative procedures (n = 1,051) were mostly women (n = 923; 88%) because most interventions were performed owing to pregnancy-related complications (n = 747; 67%) or gynecologic conditions (n = 114; 10%). General operative interventions registered included herniorrhaphy (n = 103; 9%), exploratory laparotomy (n = 60; 5%), and appendectomy (n = 31; 3%). The POMR for major operative procedures was 1% (n = 14) and was greatest after exploratory laparotomy (13%; n = 8) and caesarean delivery (1%; n = 4). Most deaths (n = 16) were a result of sepsis (n = 10–11) or hemorrhage (n = 3–5). Conclusion. The volume of surgery was low relative to the size of the catchment population. The POMR was high. Exploratory laparotomy and caesarean section were identified as high-risk procedures. Increased availability of blood, improved perioperative monitoring, and early intervention could be part of a solution to reduce the POMR. (Surgery 2015;158:7-16.) From the Department of Surgery and Perioperative Sciences,a Ume a University Hospital, Ume a, Sweden; School of Public Health, Iganga/Mayuge Health and Demographic Surveillance Site,b Iganga, Uganda; Department of Public Health Sciences,c Karolinska Institute, Solna, Sweden; Iganga General Hospital,d € e S€o dersjukhuset, Karolinska Iganga, Uganda; Department of Clinical Science and Education (KI SOS), Institute, Stockholm, Sweden; and the Department of Surgery and Perioperative Sciences,f Ume a University Hospital, Ume a , Sweden
MORE THAN 2 BILLION PEOPLE, OR 35% OF THE WORLD’S POPULATION, live in countries where only 3.5% of the world’s operative interventions are carried This work was conducted as a collaboration between Makerere University, Uganda, and Ume a University Hospital and the Karolinska Institute, Sweden. Funding was received from the Swedish Society of Medicine as well as the Golje Foundation, Sweden. The authors of this manuscript have no conflicts of interest in relation to the present manuscript or its contents. Accepted for publication March 20, 2015. Reprint requests: Jenny L€ ofgren, MD, Department of Surgery and Perioperative Sciences, Ume a University Hospital, 901 85 Ume a, Sweden. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.03.022
out.1 There is a large unmet need for surgery, and many lives could be saved by increasing the number of operative interventions.2-5 Africa carries the greatest burden of surgical disease in the world.6 Traditionally, surgery has been considered to be too costly and ineffective from a public health perspective, but cost and cost-effectiveness analyses disprove this viewpoint.7,8 Operative intervention has been identified as a neglected low-cost opportunity in both South Asia and sub-Saharan Africa, and attention to this aspect of health care in low- and middle-income countries is increasing.9-12 Reliable data on baseline surgical production are required if one is to plan for and evaluate the results of efforts to improve surgical services in lowresource settings. Comparing data on surgical output with data on the need for surgery in a SURGERY 7
8 L€o fgren et al
reference population is 1 way to measure the surgical service provision gap. Assessing postoperative outcomes will help to identify potential areas for quality improvement. One such outcome, the perioperative mortality rate (POMR), has been suggested recently as a possible indicator of accessibility to safe surgery and anesthesia.13 A number of retrospective, facility-based studies describing the caseload of surgical patients and operative procedures have been carried out in various locations in sub-Saharan Africa.14-17 In contrast, prospective studies that also include indications for operation and its outcome are scarce. The aim of the current study was to provide detailed information on operative procedures, their indications, and outcomes at 2 district hospitals in rural Uganda. METHODS Location. Data collection was carried out prospectively at the Iganga General Hospital (IGH) and the Buluba Mission Hospital (BMH) in Eastern Uganda in 2011. These hospitals were selected because they provide surgery services for the population of a so-called Health and Demographic Surveillance Site, where a study on groin hernia prevalence was carried out at the same time as the present study.4 These 2 hospitals were the primary providers of major operative procedures for a catchment population of 1.5 million.18 More specialized operative and obstetric care is provided at the nearest regional referral hospital in Jinja, located 35 km by road from the town of Iganga. This distance and the associated costs pose a barrier to the generally poor population. Private facilities that are not registered or licensed to carry out major operations may do so anyway. This study presents information from the 2 hospitals that officially carry out major operations. BMH, a nonprofit organization, does not have a defined catchment area, but basically serves the same population as the IGH. The IGH had 150 beds, 1 major and 1 minor operating room, 11 medical doctors, and 85 nurses and midwifes at the time of the study. The BMH had 120 beds with 1 operating room, 4 medical doctors, and 30 nurses and midwifes. Anesthesia in both hospitals was provided by mid-level health care providers with 3 years of training in anesthesia. The 2 hospitals had 1 gynecologist/obstetrician each. The IGH had a vacant position for a surgeon, and the BMH had a visiting surgeon once a week. All district hospitals in Uganda have positions for 5 specialist medical doctors, one of which is a surgeon. Owing to financial constraints,
Surgery July 2015
however, most hospitals do not actually recruit these specialists. In this respect, the 2 hospitals are typical of this level of hospital in Uganda. Definitions. A major operative procedure was defined as incision, excision, manipulation, or suturing of tissue that normally requires regional or general anesthesia or profound sedation to control pain.19 All other operative procedures, suitable to be undertaken under local anesthesia or nerve block, were defined as minor. Owing to overlap in treatment and anesthetic practice, all orthopedic procedures were grouped as ‘‘minor procedures.’’ The most invasive orthopedic operation in this study was insertion of a pin for traction. POMR is expressed as a percentage, and was defined as in-hospital mortality owing to any cause during surgery or within 30 days after an operative procedure divided by the number of individuals having undergone an operation. Data collection. Data were collected prospectively in 2011 using 2 pretested questionnaires, one for data collection in the operating room and one for data collection on the ward. Staff already employed at the 2 hospitals were trained to carry out data collection. Data on all major operative procedures undertaken at the 2 hospitals were collected over 4 months, and data on all minor procedures were collected simultaneously, but for 3 months only because they were much greater in number. The data collected were basic patient characteristics, type of procedure performed, indication, whether the patient was admitted or treated as an outpatient, and whether the patient was discharged, referred, or died. Complications were included, and there was also space to fill any further comments or information. J.L. and J.M. carried out quality control of the data collected. Forms filled in the operation room were crosschecked with the operating room registers and the forms filled in on the ward. If required, clinical notes from the wards and registers at the time of admission were checked to correct any inconsistency. Additional data on service provision at the 2 hospitals were retrieved from the hospitals’ annual reports. Ethical considerations. Ethical approval for the study was granted by the Institutional Review Board at School of Biomedical Sciences, Makerere University, as well as the Uganda National Council for Science and Technology. Approval for data collection was granted by the medical superintendent at each hospital. Data analysis. Double entry of the data was practiced to minimize errors. Basic statistical
L€o fgren et al 9
Surgery Volume 158, Number 1
calculations were performed using Excel 2013 (Microsoft Corporation, Redmond, WA).
RESULTS A total of 2,790 operative interventions on 2,701 patients were recorded. Of these, 1,051 patients underwent a major operative procedure. This corresponds with an annual rate of major surgery of 225 procedures per 100,000 in the catchment area population. Most patients were operated at the IGH (major, 749 [71%]; minor, 1,558 [94%]). Patients undergoing major operative procedures were mostly women (n = 913 [87%]) whereas the majority of patients undergoing minor operative procedures were men (n = 1,293 [78%]). Patients who had major operative procedures were older (mean age, 28.5 years) than those undergoing minor operative procedures (mean age, 18.4 years; P < .001). Almost all major procedures were carried out by general medical doctors. Only 17% (n = 180) were operated by either a qualified surgeon or gynecologist/obstetrician. Most minor procedures were performed by nurses, midwifes, and clinical officers. Clinical officers are midlevel providers with 3 years of clinical training. Table I shows the characteristics of the patients and the procedures carried out. Table II shows the numbers and proportions of major and minor operative procedures recorded. Table III shows the indications for the most commonly performed major operative procedures. Male circumcision, the most commonly performed minor operative procedure, was done for primary prevention of human immunodeficiency virus (HIV) in >99% of cases. The indications for the remaining minor procedures were standard. The most invasive orthopedic procedure was insertion of a traction pin in the tibial tuberosity in femur fractures. The in-hospital POMR was 0.6% (n = 16). Deaths occurred in both hospitals (13 in IGH and 3 in BMH). As shown in Table IV, death was most commonly seen after exploratory laparotomy (n = 8 [13%]) and caesarean delivery (n = 4 [0.8%]). Each death is described in detail in Table IV. Of 510 babies delivered via the 496 caesarean deliveries, 59 (12%) were either stillborn or died before leaving hospital. Hospital data revealed that overall maternal mortality in the 2 hospitals between July 2010 and June 2011 was 43 of 4,808 deliveries (0.9%) at the IGH and 10 of 1,042 deliveries (1%) at the BMH. Of the patients who died of sepsis after exploratory laparotomy, all but one had been prescribed
Table I. Characteristics of patients undergoing major and minor operative procedures at Iganga General Hospital and Buluba Mission Hospital
Characteristic
Major operative Minor operative procedures procedures (n = 1,051)* (n = 1,650)y
Female sex, n (%) 914 Age Mean, y (SD) 28.5 Children (US$1,000 per capita, confirming the grossly uneven global distribution of access to surgery.1,17 The annualized caesarean delivery rate of 99.2 per 100,000 persons per year suggests that services have increased since the 1980s when rates were reported to be around 25 per 100,000 per year in rural settings in Eastern Africa. The crude rate of
10 L€o fgren et al
Surgery July 2015
Table II. Distribution and rates of major operative procedures and distribution of minor operative procedures at Iganga General Hospital and Buluba Mission Hospital* Major surgery (n = 1,124)y Intervention Caesarean delivery Evacuation of uterus Herniorrhaphy (all kinds of abdominal wall hernia) Explorative laparotomy Hysterectomy Salpingo/oophorectomy or removal of tubal/ovarian mass Appendectomy Tubal ligation Repair of cleft lip or palate* Laparotomy for ectopic pregnancy Other Total
n (%)
Minor surgery (n = 1,666)z Rate per 100,000 persons per year
496 (44.1) 244 (21.7) 103 (9.2)
99.2 48.8 20.6
60 (5.3) 54 (4.8)
12.0 10.8
36 (3.2)
7.2
31 24 21 10
(2.8) (2.1) (1.9) (0.9)
6.2 4.8 4.2 2.0
45 (4.0) 1,124 (100)
9.0 224.8
Intervention Circumcision Repair of cuts and tears Incision and drainage of abscess Reduction of fracture Excision of granuloma of the umbilical chord Excision of mass Incision of tongue tie Reduction of dislocated joint Biopsy Removal of foreign body/implant Other Total
n (%) 768 (46.1) 293 (17.6) 238 (14.3) 138 (8.3) 36 (2.2) 35 (2.1) 22 18 16 15
(1.3) (1.1) (1.0) (0.9)
87 (5.2) 1,666 (100)
*A surgical camp for cleft lip and palate was undertaken by a plastic surgeon during the time of the study. It occurs at regular intervals, usually twice per year, but cleft palate repair is not a routine intervention provided by the hospitals at other times. yData on major operative procedures were collected over 4 consecutive months. zData on major minor operative procedures were collected over 3 consecutive months. Minor procedures are also undertaken in lower level health care facilities; therefore, reliable rates could not be generated. The number of surgeries exceeds the number of patients as some patients underwent >1 intervention at the same time. Total is not exactly 100% owing to round off error.
herniorrhaphy (21 per 100,000 persons per year), on the other hand, is lower than the 25 per 100,000 persons reported in East Africa, but similar to findings from the IGH and BMH in 2006.14,15 Considering the high prevalence of groin hernia found in the study setting, the annual hernia repair rate should ideally exceed the 130– 290 groin hernia repairs per 100,000 persons per year seen in high-income countries.4,20,21 Male circumcision has been shown to reduce the risk for men to contract HIV through sexual contact.22 Large efforts have been invested in increasing male circumcision, resulting in it being the most commonly undertaken operative procedure in this study. Incision and drainage of abscess was the third most commonly performed minor procedure. This is a simple but potentially lifesaving intervention; abscesses and soft tissue infections are the seventh leading cause of acute surgical mortality.23 Suturing of cuts and lacerations and reduction of fractures were other common minor procedures. Although conclusions about the burden of disease cannot be drawn from facility-based data, these findings indicate an high incidence of trauma in the community. Injuries contribute
greatly to the global burden of disease.24 This burden could be eased considerably through an increase in basic surgical services, such as those provided in the study setting.25 Major operative procedures on trauma patients were very rare, and many people who sustain injuries to the skull, chest, abdomen, or pelvis probably die before reaching a hospital. With future improvements in prehospital care, it is probable that the need for major surgery resources in this group of patients will increase considerably. POMR and cause of death. Saving lives is a primary goal of any health care system. POMR and deaths audit are important tools for improving the quality of service delivery. Calculating POMR identifies procedures that carry the highest mortality rates. Identifying cause of mortality will help to set priorities for both capacity improvement and future research. Auditing of POMR and complications is an inexpensive way of evaluating the quality of surgical practice. In the present study, the POMR varied with the type of operative procedure undertaken. Eight deaths were recorded after emergency exploratory laparotomy, resulting in the highest POMR (13%) in this study. The commonest indication for
L€o fgren et al 11
Surgery Volume 158, Number 1
Table III. Indications for major operative procedures Intervention Caesarean delivery (n = 496)
Evacuation (n = 244)
Hernia repair (n = 103)
Explorative laparotomy (n = 60)
Hysterectomy (n = 54)
Salpingo/oophorectomy or removal of mass in salpinx or ovaries (n = 37)
Indication Prolonged or obstructed labor Previous caesarean delivery Cephalopelvic disproportion Malpresentation Fetal distress Antepartum hemorrhage Other Incomplete abortion Bleeding Retained placenta or membranes Other Groin hernia Other abdominal wall hernia Other Gastrointestinal obstruction Peritonitis Abdominal pain Unknown Other Myoma Bleeding Prolapse Cancer (suspected or known) Other Ovarian cyst In conjunction with hysterectomy Other
n (%) 196 156 143 66 43 31 146 210 161 25 9 85 12 2 20 20 10 4 10 21 12 10 9 11 19 9 9
(39.6) (31.5) (28.8) (13.3) (8.7) (6.3) (29.4) (86.1) (66.0) (10.2) (3.7) (82.5) (11.7) (1.9) (33.3) (33.3) (16.7) (6.7) (16.7) (38.9) (22.2) (18.5) (16.7) (20.4) (52.8) (25.0) (25.0)
Totals may exceed 100% as surgery could be performed owing to >1 indication.
operative intervention was peritonitis and/or gastrointestinal obstruction, the third leading cause of death after emergency general surgery in the world.23 Emergency laparotomy is also hazardous in the high-income setting, with an overall mortality rate of 14.9% as reported in a study from the UK in 2012.26 Another study stressed the importance of timely management, involvement of senior clinicians, accessibility of suitable radiology resources, and the availability of critical care in high-risk patients.27 The timing of operative procedures could probably be improved in the study setting; many patients present late. The other 3 factors mentioned may not be correctable for many years to come. Instead, improvements not requiring a large investment should be the first priority in settings with limited resources. The use of pulse oximetry, surgical checklists, and basic hygiene routines are examples of practices that can improve operative care and outcomes.28,29 The POMR after caesarean delivery was 0.8% (n = 4), which is slightly less than the overall maternal
mortality seen at the 2 hospitals. The 4 maternal deaths in this study were a result of hemorrhage or sepsis, both common causes of maternal death in East Africa.30 All 3 deaths owing to hemorrhage were treated inadequately in terms of blood transfusion. Whether or not these deaths could have been averted is unknown, but improving the availability of blood for such patients should help to reduce maternal mortality. Most deaths in the present study were a result of sepsis. Adequate antibiotic and fluid therapy should, therefore, play a central role in reducing the risk for death after operative procedures. Of the patients who died of sepsis, all but one had been prescribed intravenous fluids and antibiotics. We cannot say, however, whether or not this treatment was given in a timely manner, in adequate doses, and with evaluation of treatment effect. Stock outs are common and patients cannot always afford to buy the medicines required. This situation, together with low staffing levels, influences adherence to recommended treatment. In addition to the low-cost improvement measures
Intervention/ Patient nr. Caesarean delivery (n = 496) 1
Deaths, Age n POMR (y)
Indication for surgery/diagnosis after surgery
Surgical planning
Post operative complication
Cause of death (as stated in questionnaire)
Days between surgery and death
Comment
Most probable cause of death
—
—
—
—
—
—
—
—
1
Grand multipara
Sepsis
Hemorrhage
0
Hemorrhage
Hemorrhage
0
Also had perineal rupture which was repaired —
4
0.8
—
—
—
Emergency
—
2
—
—
Emergency
—
3
—
—
4
—
—
8
13.3
30 Obstructed labor and sepsis 20 Eclampsia and previous caesarean section 30 Two previous caesarean sections and in labor 23 Antepartum hemorrhage — —
—
—
2
—
—
3
—
—
4
—
—
5
—
—
6 7
— —
— —
Explorative laparotomy (n = 60) 1
Emergency Hemorrhage
Emergency
—
—
0
—
—
—
—
—
—
40 Peritonitis/fecal impaction 40 Peritonitis/ intestinal obstruction 48 Peritonitis/ peritonitis
Emergency
—
—
9
—
0
Emergency Deep infection
Sepsis
6
18 Large bowel obstruction 80 Small bowel obstruction
Emergency Deep infection
Sepsis
6
Emergency Deep infection
Sepsis
4
70 Sigmoid volvulus 70 Large bowel obstruction/ intestinal obstruction
Emergency Deep infection Emergency Wound dehiscence
Emergency Severe pain
— —
7 0
12 L€o fgren et al
Table IV. Perioperative mortality rates (POMR) per 100 interventions, characteristics, and cause of death in diseased patients
Hemorrhage
Hemorrhage —
Drained 4 L of pus Sepsis from abdomen Sepsis
Also underwent sigmoid resection
Sepsis
Sepsis Also had intraabdominal adhesions
Sepsis
Sepsis Sepsis
Surgery July 2015
(continued)
Intervention/ Patient nr. 8
Herniorrhaphy for groin hernia (n = 83) 1 Repair of cleft lip/ palate (n = 21) 1
Repair of cuts and tears (n = 293) 1
Wound debridement (n = 11) 1
Surgical planning
Post operative complication
Cause of death (as stated in questionnaire)
Days between surgery and death
Comment
40 Peritonitis and small bowel obstruction/ peritonitis and intestinal obstruction — —
Emergency
—
—
0
—
—
—
—
—
—
60 Obstructed groin hernia — —
Emergency
Hematoma
—
1
—
—
—
—
—
—
Elective
Aspiration
Aspiration
0
—
—
—
—
Emergency
—
—
0
—
—
—
—
Emergency
—
—
12
Deaths, Age n POMR (y)
Indication for surgery/diagnosis after surgery
—
—
1
1.2
—
—
1
4.8
—
—
