Delivery of Operative Pediatric Surgical Care by Physicians and Non-Physician Clinicians in Malawi

Abstract

Background

Specialized pediatric surgeons are unavailable in much of sub-Saharan Africa. Delegating some surgical tasks to non-physician clinical officers can mitigate the dependence of a health system on highly skilled clinicians for specific services.

Methods

We performed a case-control study examining pediatric surgical cases over a 12 month period. Operating surgeon was categorized as physician or clinical officer. Operative acuity, surgical subspecialty, and outcome were then compared between the two groups, using physicians as the control.

Results

A total of 1186 operations were performed on 1004 pediatric patients. Mean age was 6 years (±5) and 64% of patients were male. Clinical officers performed 40% of the cases. Most general surgery, urology and congenital cases were performed by physicians, while most ENT, neurosurgery, and burn surgery cases were performed by clinical officers. Reoperation rate was higher for patients treated by clinical officers (17%) compared to physicians (7.1%), although this was attributable to multiple burn surgical procedures. Physician and clinical officer cohorts had similar complication rates (4.5% and 4.0%, respectively) and mortality rates (2.5% and 2.1%, respectively).

Discussion

Fundamental changes in health policy in Africa are imperative as a significant increase in the number of surgeons available in the near future is unlikely. Task-shifting from surgeons to clinical officers may be useful to provide coverage of basic surgical care.

Introduction

Treatment of surgical diseases in low- and middle-income (LMICs) countries has largely been neglected by the public health community in the past, despite the fact that surgical diseases constitute 11% of the global burden of disease.^1,2 This is particularly true in the world’s poorest countries, which contain 35% of the world’s population but receive only 3.5% of all surgical procedures performed worldwide.³

On average, one in every six children in sub-Saharan Africa dies before age five as a result of infectious diseases, poverty, malnutrition, traumatic injuries, and lack of perinatal care.^4–7 Comprehensive data detailing the burden of pediatric surgical disease is lacking for many sub-Saharan African countries; however, available studies suggest trauma, congenital anomalies, and surgical infections are common.^8–12 Even in countries with relatively robust general surgery practices, pediatric surgery has often received very little attention from both funding agencies and Ministries of Health.^8,13 The absence of appropriate surgical care in this environment results in many unnecessary pediatric deaths from curable surgical diseases and contributes to significant disability, ultimately compromising the quality of life of children in Africa.¹⁴

The number of pediatric surgeons serving in East, Central, and Southern Africa ranges from 1 in Malawi (population 13 million) to 25 in South Africa (population 48 million).^1,8,14,15 A significant increase in the number of surgeons, surgical specialists, and anesthesiologists in these countries is unlikely to occur in the near future.¹⁶ Therefore, national health systems must find other ways to offset the workforce shortage. One method involves mobilizing non-physician clinical officers to perform surgical and anesthetic tasks.¹

This approach, called surgical task-shifting, involves the delegation of tasks traditionally performed by surgeons and anesthesiologists to healthcare workers with lower qualifications.¹⁷ The use of clinical officers or other mid-level health workers to provide specific health interventions has commonly been used throughout sub-Saharan Africa as a strategy for expanding healthcare delivery in settings with shortages of qualified health personnel.¹⁷ Numerous studies have shown that task-shifting can be effective in the administration of ARTs,¹⁸ delivery of obstetric care,^19,20 and in surgical subspecialty care, such as ENT, orthopedics, and anesthesia.^17,21,22

In 1976, the Malawian government introduced clinical officers in an attempt to meet the health demands of its population. Clinical officers undergo formal Diploma in Clinical Medicine training at the Malawi College of Medicine, which requires 3 years of didactic education followed by a yearlong rotating clinical internship at a central or district hospital. Following internship, clinical officers are licensed to practice independently.^17,23 Currently, clinical officers provide most of the medical, orthopedic, and obstetric care at district and regional hospitals, as well as administer anesthesia at district and central hospitals.^17,23,24

The use of clinical officers in pediatric surgical care delivery in Africa is neither widespread nor formally described. Our primary objective for this study is to examine the surgical case load, complexity, and outcome of cases performed by clinical officers compared to cases performed by physicians at our institution. We hypothesize that clinical officers can perform basic pediatric surgical procedures with similar outcomes compared to physicians.

Methods

We conducted a case-control study of pediatric surgery procedures at Kamuzu Central Hospital (KCH) over one calendar year (January to December 2012). KCH is a 1000-bed tertiary hospital in Lilongwe and serves as a referral central for approximately 6 million people in the central region of Malawi. The surgical staff available at KCH during the study period included four full-time general surgery consultants, one urologist, five visiting general surgery and subspecialty consultants, and sixteen full-time clinical officers. Also operating during the study period were eleven Malawian general and orthopedic surgical residents and two American general surgery residents. The main operating theatre at KCH has four fully functional operating rooms with six clinical officer anesthetists.

All pediatric patients (≤17 years old) who underwent an operative procedure during the study period were identified in the operating theatre log book. Age, sex, preoperative diagnosis, procedure performed, operative acuity, and date of operation were recorded. Criteria for major case designation included; 1) entrance into a body cavity (cranium, thorax, abdomen), 2) need for general anesthesia, 3) soft tissue operations requiring more complex technical procedures (skin autograft, contracture release, or amputation). Criteria for minor case designation included; 1) skin, mucus membrane, or connective tissue resection only, without entrance into major body cavity, 2) local or regional anesthesia, 3) minimal degree of complexity. Patients undergoing orthopedic procedures were excluded. Patients identified in the operative log were matched to their respective entries in our pediatric outcomes database to obtain primary outcome measures including length of stay, complications requiring reoperation, and survival to discharge.

The non-physician healthcare worker in our context may be called non-physician clinicians, mid-level providers, or clinical officers. We use the term clinical officer, as this is the designation commonly used in Malawi. Cadre of operative surgeon was determined from the operative log. Procedures performed by clinical officers, with or without a surgical resident, were listed as clinical officer (CO) cases. Procedures performed by consultant-surgeons, with or without a clinical officer and/or surgical resident, or by surgical residents alone were listed as physician (MD) cases. Procedures performed by physicians were used as the control group.

Univariate and bivariate analysis was conducted to describe the pediatric surgical population and compare the outcome and operative intervention of procedures performed by physicians and clinical officers. All statistical analysis was performed using StataSE 12, Stata Corp LP, College Station, TX. Both the University of North Carolina Institutional Review Board and the National Health Sciences Research Committee of Malawi (NHSRC) approved this study.

Results

A total of 1186 operative procedures were performed on 1004 pediatric patients during the study period. Clinical officers performed 40% of all cases. The majority of cases were completed under general anesthesia (87%). Fifty-seven percent of pediatric surgical operations were major cases and approximately one-third were performed as emergencies [Table 1]. The average age of the children was 5.7 years (± 5.2), and 64% were male [Table 2].

Table 1.

Operative Characteristics for all Surgeries Performed (N=1186)

Total Operative cases	N (%*)
Anesthesia Type
General	1029 (87%)
Local/Regional	12 (1%)
Case Urgency
Emergent	380 (32%)
Elective	787 (66%)
Case Complexity
Major	672 (57%)
Minor	368 (31%)
Operating surgeon
MD	566 (48%)
CO	475 (40%)

^*Percent may not add up to 100% due to missing data

Table 2.

Patient and Operative Characteristics by Primary Operating Surgeon

Patient Characteristics	Mean (± SD) or N (%)			p-value*
	Overall (n=1004)	MD (n=507)	CO (n=378)
Age (years)	5.7 (± 5.2)	6.1 (± 5.3)	4.4 (± 4.5)	<0.001
Age Category				<0.001
Neonatal (>1 month)	66 (6.6%)	45 (8.9%)	15 (4.0%)
Infant (1 month-1 year)	160 (16%)	43 (8.5%)	106 (28%)
Child (1–10 years)	561 (56%)	298 (59%)	206 (55%)
Adolescent (11–17 years)	215 (21%)	120 (24%)	51 (14%)
Sex				0.004
Male	619 (64%)	330 (69%)	219 (60%)
Female	341 (36%)	145 (31%)	147 (40%)
Complications	42 (4.2%)	23 (4.5%)	15 (4.0%)	0.7
Reoperation	113 (11%)	36 (7.1%)	64 (17%)	<0.001
Length of Stay (days)	16 (±30)	10 (±30)	24 (±29)	<0.001
Outcome				0.8
Lived	494 (98%)	195 (98%)	184 (98%)
Died	12 (2.4%)	5 (2.5%)	4 (2.1%)
Unknown	498	307	190

^*Tests based on Pearson’s Chi-square for categorical variables and t-tests for continuous variables

General and congenital surgeries were the most common specialties [Figure 1]. Congenital pathology resulted in 23% of all cases. Gastrointestinal general surgery, soft tissue excision, trauma and burns, urology, neurosurgery, and ENT surgery made up roughly equal proportions (between 10–14% each), while surgery for neoplastic disorders was relatively rare [Figure 2]. The most common procedures were hernia repair, general surgery laparotomy, and biopsy or excision of tissue masses. These procedures were more commonly performed by physicians [Figure 3]. Burn surgery, neurosurgery, and ENT procedures were more commonly performed by clinical officers [Figure 1]. No full-time ENT surgeon was available at KCH during the study period, although visiting surgeons were available intermittently. One neurosurgical consultant was available for complex neurologic cases, although the majority of the ventriculoperitoneal (VP) shunt cases were performed by a clinical officer [Figure 3]. One urologist was available and performed the majority of the pediatric urology cases [Figure 1].

Figure 1.

Number of Procedures Performed by Physicians and Clinical Officers (By Surgical Subspecialty)

Figure 2.

Total Operative Cases (By Surgical Subtype)

Figure 3.

Common Procedures Performed by Physicians and Clinical Officers

Patients initially operated on by clinical officers were more likely to undergo multiple procedures and have a longer hospital length of stay; however, many of these patients were burn victims who underwent multiple procedures for excision and grafting [Tables 2, 3]. After excluding burn cases, the reoperation rate and hospital length of stay did not differ significantly between physician and clinical officer cohorts.

Table 3.

Characteristics of Re-operated Cases based on Cadre of Primary Surgeon

Original Diagnosis	N	Number of Surgeries	Type of Reoperation
MD
Soft Tissue Mass	2	2	1. Biopsy 2. Debulking or Resection
Wounds	4	2–3	Dressing change, repeat I&D+, pack removal
Foreign Body	1	2	1. Bronchoscopy 2. Thoracotomy
Urinary Incontinence	1	2	1. Cystoscopy 2. Bilateral Ureterostomy
Urethral Valves	2	2	Cystostomy, Dilatation, Cystoscopy
Head Injury	1	2	1. Craniotomy 2. Tracheostomy
Abdominal Wall Abscess	1	2	1. I&D 2. STSG
Rectofourchette Fistula	1	2	Staged Anoplasty
Rectovesical abscess	1	2	Staged Anoplasty
CO
Burns	51	2–9	STSG*, debridement, contracture release, disarticulation/amputation
Colostomy	1	2	Colostomy Reversal
Wounds	3	2	Dressing change, STSG*, Colostomy
Respiratory Papilloma	2	2–3	Re-excision, Tracheostomy, Decannulation
Laryngoscopy	1	2	Laryngoscopy following Tracheostomy

^*STSG: split thickness skin graft

⁺I&D: incision and drainage

The overall complication rate was 4.2%, and complication rates were similar between physicians and clinical officers, even when stratified by case complexity [Table 4]. The most common complication requiring reoperation was prolapsed colostomy, which occurred in 7 patients during the study period [Table 5]. VP shunt complications and enterocutaneous or vesicocutaneous fistulae occurred in 6 patients each and required an additional two to four surgical procedures [Table 5]. Three patients had multiple complications during the study period, requiring up to five separate surgeries [Table 5].

Table 4.

Complication Rate by Case Complexity of Initial Operative Procedure for Physician and Clinical Officer Cohorts

Case Complexity†	N (%)			p-value*
	Overall (n=883)	MD (n=505)	CO (n=378)
Major †	566 (64%)	302 (53%)	264 (47%)
Complication Rate	33 (5.8%)	16 (5.3%)	14 (5.3%)	0.998
Minor †	317 (36%)	203 (64%)	114 (36%)
Complication Rate	7 (2.2%)	6 (3.0%)	1 (0.9%)	0.2

^*Tests based on Pearson’s Chi-square.

^†Percentages reflect row percent (i.e. percent of major cases performed by physicians vs. clinical officers)

Table 5.

Description of Re-operations for Surgical Complications

Age Sex	Diagnosis	Procedure	Complication	Initial Surgeon	Total Surgeries	Outcome
4 d F	Imperforate anus	Cutback	Anal Stenosis	CO	3
21 d	Imperforate anus	Cutback	Anal Stenosis	MD	3
3 d M	Imperforate anus	Colostomy	Prolapse	MD	2	Survived
2 m M	Hirschsprung’s	Colostomy	Prolapse	MD	2	Survived
9 m F	Hirschsprung’s	Colostomy	Prolapse	MD	2	Survived
3 y M	Hirschsprung’s	Colostomy	Prolapse	CO	2	Survived
4 m M	Prolapsed Colostomy	Reduction	Prolapse	MD	2
3 m F	Bowel Obstruction	Colostomy	Prolapse	CO	3
4 y M*	Perineal Injury	Colostomy	Prolapse	MD	5	Survived
8 d	Ambiguous Genitalia	Colostomy	Obstructed Colostomy	MD	2
3 d F*	GI Perforation	Repair	Dehiscence	MD	3
9 y F	GI Perforation	Repair	Dehiscence	MD	2
21 d M	Umbilical Hernia	Hernia repair	Dehiscence		2	Survived
3 d M	Imperforate anus	Colostomy	Dehiscence	CO	2
17 y M	Bowel Obstruction	Bowel resection	Bowel Obstruction		2	Survived
3 m M	Intussusception	Reduction	Bowel Obstruction	CO	2
16 y M	Hirschsprung’s	Colostomy Reversal	Megacolon		3	Survived
9 y M	Peritonitis	Repair Perforation	Peritonitis	MD	2	Survived
11 y F	Hirschsprung’s	Colostomy	Peritonitis	MD	2
12 y M	Peritonitis	Laparotomy	GI Perforation	CO	2
16 y M*	GI Perforation	Repair	GI Perforation	MD	5	Survived
7y M	Incarcerated Hernia	Herniotomy	GI Perforation	CO	2
3 y M	Peritonitis	Laparotomy	EC Fistula	CO	3
10 y M	GI Perforation	Bowel Resection	EC Fistula	MD	2
11 y M	Ileostomy	Reversal	EC Fistula	MD	2
3 y M	Umbilical Hernia	Hernia repair	Seroma	MD	2
2 m M	Hydrocephalus	VP Shunt	Blocked shunt	CO	3
5 m F	Hydrocephalus	VP Shunt	Blocked shunt	CO	2
8 m F	Hydrocephalus	VP Shunt	Migrated shunt	CO	2
5 m F	Hydrocephalus	VP Shunt	Exposed shunt	CO	3
3 m M	Hydrocephalus	VP Shunt	Migrated shunt	CO	2
5 m M	Hydrocephalus	VP Shunt	Blocked shunt	CO	2	Survived
6 y M	Neck trauma	Tracheostomy	Tracheal stenosis	MD	3	Survived
5 y M	Prolonged intubation	Tracheostomy	Tracheal stenosis	MD	2
17 y M	Oropharyngeal Tumor	EUA	AW Obstruction	CO	2
10 y F	Hydronephrosis	Nephrectomy	Renal Failure	MD	3
17 m M	Genitalia injury	Circumcision	Revision	CO	2	Survived
10 y M	Urethral dysfunction	Dilatation	Urethral Stenosis	MD	3	Survived
10 y M	Bladder tumor	Resection	VC Fistula		2	Survived
3 y M	Urethral Valves	Destruction	VC Fistula	MD	2
10 y M	Rectourethral Fistula	Repair	VC Fistula	MD	4

^*Multiple complications including: bowel perforation, bowel obstruction, colostomy prolapse, and dehiscence

Approximately 50% of patients in each subspecialty had missing outcome data, with the exception of burn cases (7% missing outcomes). Patients with missing outcome data were generally younger (neonates or infants) and were often operated on by physicians for general surgery or congenital pathology. Of note, 26 patients with complications requiring reoperation were missing outcome data. Of the patients with known outcomes, the overall mortality was 2.4%, and mortality did not differ between the physician and clinical officer cohorts [Table 2].

Discussion

In sub-Saharan Africa, a critical shortage of human resources occurs across the entire spectrum of pediatric surgical care, including nursing, radiology, anesthesiology, and pathology.²⁵ The high cost of training physicians, combined with poor healthcare infrastructure, high surgical volume, and high rates of emigration of health professionals, result in severe workforce shortages in most African countries.^{5,7,23,24,26–28}

Surgical task-shifting represents a radical departure from traditional delivery models that depend on surgical specialists and could make a major contribution to expanding access to surgical services, especially among poor and marginalized populations. Previous studies have demonstrated that clinical officers can perform emergency obstetric, trauma-related, and general surgery procedures safely and with similar outcomes compared to physicians.^19,20,22,23 Our findings support prior studies demonstrating that task-shifting can be safe and feasible in a well-supervised environment. In particular, this is the only study that has demonstrated the effectiveness of task-shifting within a pediatric surgical population.

Within our institution, clinical officers perform the majority of pediatric burn surgery, VP shunt placement, and foreign body removal cases. Clinical officers also commonly perform major general surgery cases such as laparotomy, appendectomy, and incarcerated hernia repair, as well as minor procedures including biopsy and incision and drainage. Physicians still perform the majority of the general surgery and trauma laparotomies, congenital and reconstructive cases, and elective procedures. Although clinical officers can practice independently or with a surgical resident at our institution, consultant physicians are generally available for back-up for complex general surgery or subspecialty cases.

Pediatric surgical diseases in Africa present multiple challenges to clinicians, namely, high patient volume, delayed presentation and advanced pathology, limited resources, and lack of trained general and pediatric surgeons.⁸ Most pediatric surgeons in sub-Saharan Africa practice in large tertiary centers concentrated in major cities.^9,14 This has serious implications for the effective delivery and utilization of pediatric surgical services in the continent, as more than two-thirds of the population resides in rural areas.¹⁶ Children with surgical diseases may travel long distances to reach an urban hospital, which may be partially responsible for the delayed presentation and advanced pathology often seen in patients presenting for specialist pediatric surgical consultations.^8,29

Task-shifting could be used to relieve the surgical burden of disease in one of two ways. In tertiary centers with available surgical subspecialists, clinical officers could perform minor pediatric surgical procedures, leaving consultant-surgeons free to perform major cases. Alternatively, in the absence of subspecialists, well-trained clinical officers could serve as the first line of treatment for common pediatric surgical diseases at the district level. At our institution, senior clinical officers successfully performed technically complex cases such as exploratory laparotomy with bowel resection and VP shunt placement. Clinical officers of this caliber could be used to staff district health centers and thereby decrease the referral rate to tertiary hospitals.⁹ Given the severe shortage of pediatric surgeons in most sub-Saharan African countries, task-shifting some common surgical care to clinical officers may have the greatest potential to provide coverage in underserved rural areas.¹

The use of non-physician healthcare workers has been practiced not only in LMICs, but in high-income countries (HICs) as well. In HICs, workforce shortages generally reflect misdistribution rather than an absolute shortage of physicians. Attracting and retaining physicians, particularly specialists, to rural areas has often presented a challenge. As a result, the medical environment is evolving to include non-physician providers as a way to meet healthcare needs. As physicians become more super-specialized, many medical practices in HICs use independently practicing physician assistants or advanced nurse practitioners as the first step in preventive and primary care. Surgical services have been slow to adopt non-physician healthcare workers in HICs, in part due to patient expectations and medico-legal concerns. However, non-physicians have been successfully introduced in specialties such anesthesiology under the supervision of attending physicians.

The results of our study suggest that clinical officers can be trained to manage many common pediatric surgical cases safely. However, this should in no way imply that pediatric surgeons can or should be entirely replaced by clinical officers. Traditional medical school and surgical residency, although time-consuming and expensive, provides a level of surgical decision-making and medical knowledge that is beyond the scope of a clinical officer training program. Many major pediatric surgical cases, particularly congenital and neoplastic conditions, require complex perioperative management and technical skill and should be managed instead by pediatric surgeons. However, given the current workforce constraints in most LMICs, adequate numbers of physicians are unlikely to be available in the near future and clinical officers could assist greatly in the management of common, uncomplicated surgical disease. A dual approach of training both physicians and clinical officers must be implemented in LMICs. A physician-only workforce is both impractical and unachievable in most LMICs and clinical officers can play a valuable role in the healthcare workforce.

Implementing task-shifting within surgical specialties will present several challenges to ensure that patient care is not compromised.^18,23 Clinical officers and other mid-level healthcare providers must be appropriately trained and qualified for the new tasks they will be asked to undertake. This may require broadening the clinical internship to include a rotation at a tertiary center for surgical training. Continuing education and training workshops, along with required credentialing, may help ensure that the clinical officers have knowledge and skills necessary to triage and manage patients appropriately.³⁰ Recent guidelines from the World Health Organization (WHO) emphasize that task-shifting must be adopted within a broader strategy of strengthening health systems, including introducing mechanisms and research to ensure that quality of care is not compromised.³¹ Ultimately, the licensing and supervision of clinical officers will need to be addressed by the national regulatory and licensing bodies in a way that reflects the country’s resources and goals.

Task-shifting has been praised on several fronts for its potential to improve the skill mix of teams,^26,27 lower costs for training and remuneration, shift healthcare to cadres that are more easily retained and dispersed to rural areas,¹⁹ and support retention of physicians by reducing burnout.¹⁸ However, both physicians and clinical officers have expressed concerns with the long-term effects of task-shifting. Some physicians are resistant to changes out of fear of loss of hierarchy, concern for patient safety, loss of earnings, and the additional supervisory responsibilities required.³² Some clinical officers, particularly senior clinical officers with many years of experience, have expressed frustration with the lack of professional development. Formal recognition of clinical officers through credentialing may help to overcome the bias against task-shifting in the surgical community. Career progression is important for this cadre of healthcare workers. In Malawi, this has been accomplished by creating a pathway for clinical officers to enroll in medical school and graduate within a shorter amount of time. This system has the two-fold benefit of training clinical officers to fill the immediate need for healthcare workers, while allowing experienced clinical officers to transition to physicians as the health systems develop.

The main limitation of this study is the quality of data collection, particularly with outcome data. In our setting, we are unable to follow patients after they are discharged from the hospital; therefore we limited our outcome analysis to survival until discharge rather than long-term survival. Subjects were initially identified in the operative log and then matched to corresponding entries in the inpatient registries. However, the inpatient registry in the pediatric department at our institution is limited and many patients were not captured. Many subspecialty cases (Neurosurgery, ENT, and Urology) are performed on an outpatient or short-stay basis, and these patients are not captured in the inpatient registries, although they likely survived to discharge. However, approximately 50% of General and Congenital Surgery patients were also missing outcome data, including 26 patients with complications requiring reoperation. This primarily reflects a failure in data collection at our institution and marks an area for improvement in the future. Given these limitations, our estimation of overall mortality in both the physician and clinical officer cohorts is limited and we concentrated our attention on differences in complication rates between the two cohorts.

For the purpose of this study, we focused on complications requiring reoperation, as these were captured in the operative log. This study demonstrates that clinical officers and physicians have similar rates of complications resulting in reoperation, even when performing major surgical cases. Non-operative complications are inconsistently diagnosed and recorded in the medical record in our setting; therefore, we are unable to estimate the burden of non-operative complications in the physician and clinical officer cohort. We hope to be able to strengthen the inpatient registry at this institution in the coming years to better assess total complication rates and overall mortality.

Conclusion

Task-shifting in pediatric surgery is feasible for common pediatric surgical procedures within a tertiary hospital with considerable oversight and supervision. We believe that by reorganizing the surgical workforce in this way, task-shifting can make more efficient use of the scarce surgical human resources currently available in sub-Saharan Africa.