Background: Craniopharyngiomas account for 6-10% of childhood brain tumours and are the third most common intracranial tumours of childhood. Despite their benign histological appearance, they are often associated with a poor outcome and have significant associated morbidity. Aim: To analyse the data of children with craniopharyngiomas at Red Cross Children's Hospital and Groote Schuur Hospital from 1976 to 2004 with respect to age at presentation, presenting symptoms and preoperative and post-operative endocrine dysfunction. Patients and methods: The records of 45 children aged between 9 months and 13 years were reviewed. The majority of children in the study were aged between 5 and 10 years. There was a considerable delay in the diagnosis in most cases especially in areas outside of Cape Town. Pre-operative tests prior to 2000 were incomplete but have improved since then. Twenty nine percent of our patients had a near total excision of the craniopharyngioma and 29% had partial excisions necessitating adjuvant radiotherapy. The endocrinological, neuro imaging and presenting symptoms were analysed and post-operative tests were reviewed. Where possible neuropsychiatric assessments were accessed as well in order to assess long term neurocognitive deficits Results: The age of presentation of craniopharyngiomas in our group of patients was much younger than in other studies with the largest group of affected children being between 5 and 10 years of age. The most common presenting symptoms in this group were headaches (62%) and visual disturbances (57%). Sixty four percent of the children had preoperative endocrine testing for pituitary dysfunction of which 59% were ACTH and TSH deficient and only 38% were growth hormone deficient. After surgery multiple endocrinopathies were universal with 56% of children having pan hypopituitarism at follow up. There was no reversal of pre-existing hormone deficits after surgery. The management of craniopharyngiomas remains controversial. Twenty nine percent of our patients had a near total excision and 29 % had a partial excision combined with radiotherapy with a further 6 patients receiving intratumoral bleomycin for recurrences. Post operatively the majority of children were on replacement therapy but only 6 patients (14%) received human growth hormone or are receiving human growth hormone currently due to a lack of funds to provide human growth hormone for children with growth hormone deficiency. Seventy percent of the children had central diabetes insipidus post operatively reflecting posterior pituitary damage. Our overall survival rate from 1976 to 2004 was 91 % with a mortality rate of 13%. Most of the deaths were attributed to recurrences of the craniopharyngioma and the complications of bleomycin treatment. Conclusion: Craniopharyngiomas remain tumours associated with significant morbidity. Total excision of the tumour remains a favourable option but the proximity of the craniopharyngioma to the pituitary gland and optic tracts results in devastating sequelae. Although the medical and surgical management of craniopharyngiomas has improved, a significant number of patients had endocrine sequelae. The management of craniopharyngiomas in an African context compares favourably internationally but with limited resources especially with regard to growth hormone replacement, a large proportion of our children are not receiving optimal treatment.

Background: The South African Road to Health Card (RTHC) is a parent-held personal child health record used by health professionals and parents to monitor the health and development of a child. Low possession and retention of the RTHC by parents, as well as inadequate use of RTHC by health professionals, have lessened its efficacy and purpose. Objectives: This study aimed to describe the level of possession of the RTHC by a sample of caregivers of patients admitted to Red Cross War Memorial Children's Hospital (RCWMCH), and to determine the extent and accuracy of doctors' transfer of clinical information between the RTHC and hospital records. Methods: A cross-sectional and analytical study was conducted in four general paediatric wards at RCWMCH. Consecutive patients were enrolled at discharge during office hours over a six week period. Data were extracted from a photograph of the RTHC and the participant's hospital record. Both the 1995 Chart version and the 2009 Booklet version of the RTHC were included in the study. The presence or absence of selected items of information on the RTHC and the hospital record were recorded - the primary outcome was the transfer of the specified information between records. The outcomes are presented as proportions, with 95% confidence intervals (CI).

Background This study attempted to identify as far as possible the extent of the children’s nursing workforce in five selected countries in the sub-Saharan African region. Strengthening children’s nursing training has been recommended as a primary strategy to reduce the underfive mortality rate in African nations, including South Africa and Malawi. The current level of data monitoring capacity worldwide means that it is not possible to disaggregate the children’s nursing workforce in countries in the World Health Organisation African Region from the data provided by the WHO Global Atlas of the Health Workforce database. Yet developing an accurate depiction of the specialist children’s nursing workforce is a necessary step towards optimizing children’s health service delivery. Methods In attempting to respond to this need, this study adheres to a collaborative research philosophy, using a convergent parallel mixed methods design, incorporating a scoping documentary review, together with quantitative (surveys and case study compilation) and qualitative (interview) components collected independently and then integrated during analysis and interpretation, to generate data addressing three related questions: how many children’s nurses are believed to be in practice nationally; how many such nurses are recorded on the nursing register nationally; and how many children’s nurses are being produced through training. Results Findings suggest there are approximately 3 728 children’s nurses across the five countries in this study. A combined total of 260 children’s nurses are produced through training each year across the five countries on average. Survey responses, interview data and content analysis of items identified through the scoping review suggest that adequate information regarding the children’s nursing workforce is not currently available to inform decision-making. Conclusion In conclusion, it is hoped that the data generated might contribute towards identifying the size of the children’s nursing workforce, as a first step towards identifying what would represent a viable and sustainable regional children’s nursing workforce for the future.

Aim: Noninvasive Continuous Positive Airway Pressure (nCPAP) and High Flow Nasal Cannula oxygen therapy (HFNC) are non-invasive ventilation (NIV) modalities appropriate for children in developing countries. There is minimal literature describing nCPAP and HFNC use in children with respiratory compromise secondary to non-pulmonary disease. This study aimed to describe the characteristics and outcomes of all children without primary lung pathology, who received nCPAP and HFNC during their admission to Red Cross War Memorial Children's Hospital, Cape Town, South Africa. Methods: This was a prospective observational study of routinely collected data, between August 2015 and January 2016. Primary and secondary outcome measures were NIV failure (progression to intubation and invasive ventilation) and Paediatric Intensive Care Unit (PICU) admission respectively. Comparative statistics were conducted using Mann-Whitney U tests. Data significantly associated with the primary and secondary outcomes on univariate analysis were entered into backward stepwise logistic regression models to determine independent predictive factors. Results: There were 31 cases of nCPAP and one case of HFNC use in 31 patients (median (IQR) age 3.5 (1.8 – 7.6) months. The majority (n=23; 71.9%) presented with primary diarrhoeal disease. There were two deaths (6.5%), 17 (53.1%) Paediatric Intensive Care (PICU) admissions, and five (15.6%) cases received invasive ventilation (NIV failure). Median (IQR) duration of hospital stay was 11.50 (6.0 – 17.5) days. Patients who failed NIV had lower admission SaO2 than those without treatment failure (95 (95-99)% vs 100 (100-100)%; p = 0.03). On multiple logistic regression, lower temperature (OR 0.19; 95% CI 0.05 – 0.78; p = 0.02) and receiving inotropes in the emergency setting, (OR 23.05; 95% CI 1.64 – 325.06; p = 0.02) were independently associated with PICU admission. Conclusions: nCPAP is used clinically for the management of children with respiratory compromise secondary to non-pulmonary illnesses, particularly diarrhoeal disease. Larger controlled clinical studies are needed to determine the effectiveness and utility of nCPAP in this population. HFNC was not commonly used, and this modality requires further investigation in this population.

In the past two decades major focus has been placed on child's rights, survival and development. This has occurred both on an international level and, since the inception of the Government of National Unity (GNU) in South Africa, nationally, provincially and locally through various development programmes. Development in a child concerns the sequential acquisition of cognitive, motor and social skills. Adverse biological and environmental factors will have a major negative impact on a child's development. It is therefore crucial that children affected by these adverse factors are identified as early as possible, to prevent disability or facilitate intervention to ensure that they reach their maximum developmental potential. The global prevalence for developmental disability (DD) is 7-10% of the population. The national prevalence for South African children is not known but a recent Disability Survey by the Department of Health in 1998 quotes 5. 7% - 6.1 % for the overall population. Developmental disability is therefore an important priority to be addressed, especially at the primary health care level. This has been highlighted in the White Paper on Integrated National Disability Strategy of the GNU in 1997. The efficacy of intervention for developmental disabilities is debated. Scientifically this efficacy has not been proven unequivocally but there is consensus from parents, professionals and advocacy groups that early intervention is beneficial. Taken that early identification is essential, screening is the ideal method in the South African context for detection of developmental disability. There are numerous screening tools used for this purpose which need to fulfil certain criteria to ensure effectiveness. Important issues around existing developmental screening tools have been: time taken to administer the test; reliability, validity and sensitivity; measurement methods; ease of administering; and applicable age range. Questions which record parents' or caregivers' observations of their child's developmental skills have been used as effective screening tools for developmental disability. A large study which assessed prevalence of disability using a "Ten Question" format has been used extensively in Bangladesh and found to be effective. A 10-question tool was developed to screen for developmental disabilities in 9-month-old infants, the age of attendance at clinics for measles immunisation. The questions focused on the four areas traditionally tested in developmental assessments, namely speech and hearing, gross and fine motor, vision and personal/social functions. Other factors such as weight, head circumference, dysmorphic features and chronic illness were documented concurrently. The 9-month screening tool in this study was tested at Site B Clinic (Khayelitsha) and Eastridge Clinic (Mitchells Plain) and extended to high risk clinics such as the Neonatal Elevated Risk, Cerebral Palsy and Developmental Clinics of Red Cross War Memorial Children's Hospital. This was a prospective observational study. 235 9-month-old infants were seen between the period September 1996 to January 1997. The interviews were conducted in Xhosa (53%), Afrikaans (28%) and English (19%). The questionnaire was administered to the parent or caregiver of each infant. This was followed by an examination and neurodevelopmental assessment of the infant. On assessment 192 infants (82%) had no disability, 39 (16%) had developmental disability and 4 (2%) were at risk for motor disability. A further analysis of the 39 infants with developmental disability revealed that 25 (64%) were language impaired only, 2 (5%) were motor disabled and 12 (31%) were globally disabled (i.e. in both language and motor areas). Males and females were equally affected. The results showed that the 9-month screening questionnaire is a valid tool for the screening for developmental disability. The overall sensitivity of the questionnaire was 97.7% (95% Confidence interval [CI] of 95.8 - 99.6) and specificity 88.5% (95% CI of 84.4 - 92.6). The positive predictive value was 65.6% and negative predictive value 99.4%. Only one infant who was at risk for motor disability was missed by the questionnaire. On evaluation of the questions in specific areas (i.e. speech, hearing, motor, vision, squint), the numbers assessed were too small for meaningful analysis. As personal and social development involves motor, language and hearing skills, this area was not used on its own in the overall assessment of the infant, but those questions were incorporated into other areas of the screen. Of concern was that a number of infants screened positive for DD on certain questions (e.g. language), but on assessment were found to have disability i n a different developmental area (e.g. motor). In these cases, the questionnaire identified the children with disability - but for the incorrect reason. Certain issues would need to be addressed prior to implementation of the screening tool to ensure an adequate, appropriate screening procedure and prevention of falsely raising the expectations of parents. This involves personnel training, monitoring of screening, referral mechanisms, intervention once DD is identified and involvement of the family as a partner in the process.

Background: Dehydration due to acute gastroenteritis (AGE) remains a leading cause of child death worldwide. The primary treatment is enteral rehydration. Children who fail a trial of oral fluids require rehydration in hospital, preferably via nasogastric tube. Traditionally, children have been rehydrated over 24 hours; 'Standard Rehydration' (SR). Most treatment guidelines now recommend 'Rapid Rehydration' (RR) over 4-6 hours. There are limited data comparing RR to SR, especially from low-resource settings. Objectives: To assess the efficacy and efficiency of RR in children with AGE in the Rehydration Unit of Red Cross War Memorial Children's Hospital, Cape Town. Methods: A retrospective cohort study was performed. The intervention cohort contained 67 children who received RR in March 2007. The control cohort contained 76 children who received SR in March 2006. The outcome measures were weight and hydration status at 4 hours and time to maximum weight to measure efficacy; and length of hospital stay (LOS) to measure efficiency. Results: Children in the intervention cohort experienced greater weight gain (p<0.01) and lower dehydration scores (p=0.01) at 4 hours. There was no difference in time to maximum weight. The LOS for the two groups were not statistically different. Conclusion: RR is an effective method of rehydrating children with AGE. In contrast to two studies in well-developed settings, reduction in LOS following RR could not be demonstrated. There is no reason not to adopt RR as the predominant rehydration method in settings such as ours. More research is required to evaluate the efficiency of RR.

Chromosome 22q11.2 deletion syndrome (22qDS) is an inherited autosomal dominant disorder. It is the second most commonly occurring syndrome, Trisomy 21 being the most common. It is the most common microdeletion syndrome. The clinical range of features with which affected individuals present is very broad and includes congenital heart disease (particularly conotruncal malformations), palatal abnormalities, characteristic facial features, and learning difficulties. In total, there are more than 180 different phenotypic features associated with 22qDS. Due to the wide variability of phenotypic features that can arise in 22qDS it is often difficult to know when to test for the syndrome. Oskarsdottir’s criteria are widely used in clinical recognition for 22qDS. However, in a study done previously at Red Cross War Memorial Children’s Hospital, this same criteria was found to only have a positive predictive value of 14% for 22qDS. This is likely due to the fact that Oskarsdottir’s study was based on a largely Caucasian study population. Our population in Southern Africa is largely non-Caucasian. Previous studies have confirmed that non-Caucasian population groups with 22qDS have different presenting phenotypic features. For this reason, in this study we sought to describe the typical phenotypic features with which children with 22qDS present in our local population in South Africa. A retrospective folder review was done of the phenotypic features of all patients who had had a FISH test done on the suspicion on 22qDS. A total of 144 patient folders were reviewed (72 patients who were FISH positive for 22qDS and 72 patients who were FISH negative and functioned as the control arm of this study). A review on the phenotypic features of children with 22qDS revealed the most common presentation to be congenital heart disease (44%), failure to thrive (33%), dysmorphic features (32%) and cardiac failure (25%). A positive family history was only noted in 13 patients. Of those patients with a positive family history of 22qDS, only 5% were proven FISH positive for 22qDS themselves (less than the 10% described in the literature). Younger children presented more frequently with CHD, while older children presented with developmental delay and dysmorphic features. In general, developmental delay, palatal abnormalities and feeding difficulties were less common in our study population than described in the literature. Our particular patient population presented with the following CHD: isolated VSD (46%), tetralogy of Fallot (20.8%), truncus arteriosus (14.5%), PS/pulmonary artery stenosis (20.8%) and interrupted aortic arch (6%). Interrupted aortic arch was found to be the most sensitive marker for 22qDS in children with cardiac lesions. The cardiac lesions with the highest positive predictive value for 22qDS was non-isolated VSD (54%). Dysmorphic features with the highest sensitivity for 22qDS included bulbous nose (75%), abnormal digits (64%) and posteriorly rotated ears (68%). Primary immune deficiency, thymus abnormalities, cleft palate and behavioural issues were described less in this study than previously described in the literature. In conclusion, it is clear that non-Caucasian populations have some unique phenotypic expressions of 22qDS. It is imperative that clinicians maintain a high index of suspicion for patients with 22qDS.

Background: Discharge diagnostic data from hospital administrative databases are often used to inform decisions relating to a variety of vital applications. These may include the allocation of resources, quality of care assessments, clinical research and for the formulation of healthcare policy to name a few. Having accurately coded and reliably captured discharge data for patients is of paramount importance for any hospital and health system to function efficiently. Objectives: 1. Retrospectively examine the reliability of the ICD discharge coding in the Red Cross War Memorial Children’s Hospital (RCWMCH) administrative database for primary and secondary discharge diagnoses. 2. Formulate recommendations for improvement to the current system. Methods: This study was a retrospective folder review of 450 patient admissions to the short stay and general paediatric wards at the RCWMCH between 1 August 2013 and 1 September 2014. International Classification of Diseases version 10 (ICD-10) discharge coding was completed for each admission by the principal investigator and compared with the corresponding admission data captured for each patient within the Clinicom® Health Information System. Agreement comparison was done to both four-character and three-character ICD-10 code specificity. Results: Of the initial 450 randomly selected folders, 396 (88%) were analysed during the folder review process. The median number of total diagnoses (primary diagnosis plus secondary diagnoses) coded by the principal investigator (PI) folder review was 3 with a distribution of 1 to 10 (IQR 2 - 4). The median number of total diagnoses coded in Clinicom® was 1 with a distribution of 1 to 3 (IQR 1 - 1). Agreement of primary diagnosis coding to four characters was 26.3% with slight improvement to 34.3% when assessed to three characters. Agreement for secondary diagnoses to four characters was 14.9% and 27.7% when assessed to three characters. Conclusion: Reliability of administrative discharge data from RCWMCH is poor. Inadequacies with regard to the employment of dedicated and/or adequately trained coding personnel may be significant contributors to the problem.

Background Despite improved treatment and access to care, adolescent AIDS deaths are decreasing more slowly than in any other age group. There is lack of longitudinal data around adolescent adherence and the dynamics of viraemia over time. We aimed to describe patterns of detectable viral load in a cohort of adolescents attending an antiretroviral clinic in Cape Town, South Africa. Methods We conducted a retrospective cohort study of all patients on ART aged 10-19 years. Participants were included if they underwent at least two HIV viral load (VL) measurements and attended the Groote Schuur Hospital HIV Clinic for at least 24 months between 2002 and 2016. The primary outcome was two consecutive VL >100 copies/ml, in line with the lower limit of detection of assays in use over the follow-up period. Results Of 482 screened subjects, 327 met inclusion criteria. Most subjects were vertically infected (n= 314; 96%), and 170 (52%) were male. Overall, 203 episodes of confirmed detectable VL involving 159 (49% [95% CI 43%–54%]) subjects were experienced during the follow-up period. A total of 111 (34%) subjects never experienced detectable VL, while 16 (5%) never suppressed throughout the follow-up period. Median age at first detectable VL was 14 (IQR 11-16) years. Of the 159 subjects who experienced detectable VL, 102 (64%) re-suppressed, of which 38 (37%) had a subsequent detectable VL. Six subjects had genotyped resistance to protease inhibitors. Four of these never suppressed, while two suppressed on salvage regimens. Total follow-up time was 1723 person years (PY), of which 880 (51%) were contributed by the 159 subjects who experienced detectable VL. Overall time with detectable VL was 370 PY. This comprised 22% of total follow-up time, but 42% of the follow-up time contributed by those who experienced detectable VL. The rate of detectable VL was 11.8 (95% CI 10.3–13.5) episodes per 100 PY. The risk increased by 24% for each year of increasing age (RR 1.24 [95% CI 1.17-1.31]; p< 0.0001). Neither prevalence, duration nor rate of detectable VL was influenced by gender. Conclusion Detectable VL was seen in nearly half of adolescents, with the rate increasing with age. Viraemia was not a static process, and adolescents moved in and out of this state as adolescence progressed. Further study is warranted to correlate these findings with risks and clinical outcomes.

Includes abstract. / Includes bibliographical references. / From November 2002 to December 2006, a placebo-controlled, randomized trial investigated the incidence of tuberculosis and the overall mortality in a cohort of HIV-infected children in Cape Town, South Africa. They were randomized to receive either Isoniazid Preventive Therapy (IPT) or placebo. In addition, they were randomized to receive trimethoprim/sulfamethoxaxole prophylaxis on either a daily or a three-times-per-week schedule. The aim: To describe the incidence of bacteraemia, and the spectrum of organisms cultured. To determine if there was a difference in the incidence of bacteraemia between children using Isoniazid Preventive Therapy (IPT) versus placebo; and to determine if there was a difference in the incidence of bacteraemias between the groups using daily versus thrice-weekly trimethoprim/sulphamethoxazole prophylaxis.