Enhancing Diabetes Diagnosis with AI and ML in Abuja's Pediatric Unit

Categories: Science

Essay, Pages 10 (2361 words)

Views

Introduction

Abuja is the Nation’s seat of power and it is located within the federal capital territory (FCT). The FCT has an estimated population of 1,406,239 people as of the 2006 population census and occupies a landmass of approximately 7,315 km2. It is bordered by the states of Kaduna to the northeast, Kogi to the southwest, Nasarawa to the east and south, and Niger to the west and north.80 The population is a cosmopolitan comprising of several ethnic groups in Nigeria. Abuja Municipal Area Council (AMAC) is located on the eastern part of the FCT and occupies a landmass of approximately 250 km2.

Don't use plagiarized sources. Get your custom paper on

“ Enhancing Diabetes Diagnosis with AI and ML in Abuja’s Pediatric Unit ”

Get high-quality paper

NEW! smart matching with writer

National Hospital Abuja is a government-owned tertiary referral center located in the central business district of AMAC. The hospital provides services to patients in Abuja and other nearby states. The Paediatric Department of the hospital consists of an Emergency Paediatric unit, In-Patient Paediatric Ward, Special Care Baby Unit, Neonatal Intensive Care Unit, and outpatient specialist clinics.

The Emergency Paediatric Unit is a 24-bed ward that receives emergency cases and referrals from other hospitals within and outside the state.

It provides 24 hours service and has a total of 25 nurses that runs 3 shifts with an average of 3 to 4 nurses per shift. It has a total number of 5 consultants, 6 resident doctors, and 6 house officers with at least 4 to 5 residents, 1 consultant, and 3 to 4 house officers per shift. The annual admission of children with sickle cell disease presenting with sickle cell crises is about 198 with an average monthly admission rate of 15 to 18 children.

Study Design

The proposed study will be a cross-sectional descriptive study.

Study Population

The study will involve all children aged 2-15 years with sickle cell disease admitted into the Emergency Paediatric Unit of NHA in sickle cell crises during the study duration. Sickle cell disease in these patients will be diagnosed by Haemoglobin electrophoresis.

Inclusion Criteria:

All children diagnosed to have SCD aged 2-15 years admitted for sickle cell crises
Children who gave their assent (where necessary) and also whose parents/guardians consented to the study.

Exclusion Criteria:

Children that are not in crisis.
Children with pre-existing renal disease.
Children with known co-morbid conditions such as diabetes mellitus, hypertension, HIV infection, nephrotic syndrome, obstructive uropathy, which may affect kidney function.
Children on steroids.
Adolescent female subjects menstruating at the time of the study.
Children who have a history of use of radio-opaque dye and some drugs such as nitrofurantoin, cephalexin, cephalothin, captopril, tetracycline 4 weeks earlier, as these agents decrease the reactivity of dipstick.
Parental refusal to give consent or refusal of assent by children for more than 7 years.

Sample Size Determination

Estimated sample size using the formula for calculation of sample size if the population is greater than 10000,

n = Z2pq

Where n = desired sample size if the population is more than 10,000

Z = the standard normal deviate usually set at 1.96 corresponding to the 95% confidence interval

P = the proportion in the target population estimated to have the attribute of interest.

(Prevalence of AKI during VOC = 17%) as recorded by Baddam et al 15 q = 1 – p (the proportion of the population without the attribute)

d = degree of accuracy desired set at 0.05 (5%)

n = (1.96)2 x 0.17 x 0.83

(0.05)2

n = 216.7 that is approximately 217

However, the study population is less than 10000 hence the desired study population size is calculated using the formula:

nf = n

1+ n

Where

nf = minimum sample size when the study population is less than 10000

n = minimum sample size when the study population is greater than 10000

N = the estimate of the study population size per annum (derived from the statistics department of the NHA) that is, 198

nf = 217

1 + 217

198

nf = 103.5 that is approximately 104

nf = 104. However, giving an attrition rate of 10%, n = 114

Sample size for the study = 114

Data Collection

The data will be collected by the researcher and two trained research assistants (trained by the researcher in both data and sample collection).

The records of all SCD patients will be evaluated and a pretested proforma (Appendix I) will be issued to the study population and their respective caregivers to obtain the following variables: age, sex, socioeconomic status, number and types of crisis in a year, duration of hospitalization, number of blood transfusions in the last 1 year, use of hydroxyurea of at least 3 months and above, frequency of NSAID use in the last 1 year, history of previous renal disease and use of steroids. The socioeconomic status will be assessed using Olusanya’s socio-economic classification scheme (Appendix III).

Clinical examination of the child will be done. Weight, height, pyrexia, pallor, jaundice, blood pressure, hand and foot swelling, hepatomegaly, and splenomegaly will be assessed.

Weight in kilogram (kg) will be measured with SECA model 764, Alibaba Company, Germany Body weight scale. The subjects will be weighed barefoot, wearing clothes with empty pockets. For patients that are unable to stand, the weight will be calculated by subtracting the weight of the mother from the combined weight of the child and mother (carrying the child) initially taken together. The average of two measures will be used as the weight of the patient. This type of scale has a range of 0 – 100kg with a standard error (sensitivity) of 100g. The accuracy of the scale will be confirmed before use each day and repeated at regular intervals after every 10th use.

Height in centimeters will be measured using a stadiometer attached to the SECA model 764 weighing scale from Alibaba Company Germany. The subjects will be barefooted, standing erect with the heel, back, and occiput against the vertical scale of the stadiometer. The subject’s heads will be positioned to look forward to the lower border of the eye sockets in the same horizontal plane as the external auditory meatus. The horizontal board attached to the vertical scale of the stadiometer will be lowered until it touched the subject’s head. The height will then be read to the nearest 0.1cm. The measurements will be taken twice and the average height for the subject will be noted. This SECA model 764 stadiometer from Alibaba Company, Germany has a sensitivity of 0.1cm and a precision of 1mm.

The Body Mass Index (BMI) will be calculated for each subject as weight (kg) / height2 (m2). BMI value for age and sex between the 5th percentile to less than the 85th percentile will be regarded as normal or healthy weight; between the 85th percentile to less than 95th percentile as overweight; BMI ≥ 95th percentile as obesity while BMI ˂ 5th percentile will be regarded as underweight.

Blood pressure will be measured indirectly with the use of a sphygmomanometer. A cuff (width at least 40% of the arm circumference) will be attached to a mercury sphygmomanometer and inflated around the extended arm in a sitting position. The cuff will be inflated until the radial pulse of the same hand is not felt. The stethoscope will then be placed in the cubital fossa, over the brachial artery, and the pressure released until the systolic and diastolic readings are obtained. The systolic pressure will be taken as the point at which Korotkoff sound becomes audible (phase 1) and the diastolic when it becomes muffled (phase IV). The measurement will be done twice and the average taken as the blood pressure of the subject. The child will be considered normotensive if the BP is ˂ 90th percentile, prehypertensive if ≥ 90th percentile but ˂ 95th percentile, and hypertensive if ≥ 95th percentile for sex, age, and height.

Sample Collection

Two milliliters of blood will be collected to determine the genotype of patients whose genotype has not been confirmed with haemoglobin electrophoresis machine (consort E3700 cellulose acetate system) at a PH of 8.4-8.6 using cellulose acetate paper.

The subjects will be recruited simultaneously as they presented in crises at the Emergency paediatric unit of National Hospital Abuja. 5mls of venous blood will be collected from each patient using an aseptic procedure before the commencement of treatment for the crises. 3mls of blood will be placed in a lithium heparin bottle for serum creatinine while the remaining 2mls of blood in an ethylene diamine tetra-acetic acid (EDTA) bottle for haemoglobin concentration estimation, reticulocyte count, and packed cell volume (PCV).

About 15mls of a midstream urine sample will be collected by the caregiver (or the subject for an older child) in a labeled sterile universal bottle. The caregiver/subject will be instructed on how to collect the urine sample: they should wash the head of the penis or labia with clean water, pass some amount of urine into the toilet and then place the container to collect the urine up to the 15ml mark and bring the sample to the researcher/ research assistant within 2 hours of collection (as recommended by the manufacturers of the urinalysis dipstick).

Assay Method

The sample collected in the EDTA bottle will be sent to the haematology laboratory where haemoglobin level, reticulocyte count, and PCV will be determined using an automated blood cell analyzer (Sysmex mode NK21). Anaemia in children with SCD is defined as Hb less than 11g/dl. Anaemia in children with SCD is classified as mild (9-7g/dl), moderate (˂ 7-5g/dl), and severe (˂5g/dl).

Serum creatinine will be determined by analyzing the 3mls of blood in the lithium heparin bottle using the Jaffe's method on corning colorimeter reading at 520 nanometers. The sample collected in the lithium heparin bottle will be spun at 2500 revolutions per minute (RPM) using the Eppendorf centrifuge 5810 and then loaded into the Roche Cobas c 311 autoanalyzer system which will read off the serum creatinine level. Samples will not require storage as they are going to be used within 4 hours of arriving in the chemistry laboratory. The samples will be kept at room temperature of 25oC and analyzed as soon as it gets to the chemistry laboratory.

The Jaffe's method is based on the reaction of alkaline picrate with creatinine to form an orange-red compound, the Janovsky complex.85 the intensity of the colour produced is directly proportional to the amount of creatinine present and the reading is automated. The coefficient of variation is ± 2% for within-run samples. The ideal sample bottle for serum creatinine collection is plain. However, most chemical pathology laboratories prefer blood in lithium heparin bottle because heparinized blood does not interfere with the serum creatinine determination until the concentration of heparin reaches 1g/l in the specimen and the only 250ul of serum is required.86

Urinalysis will be done using multistix 10SG dipstick (Made in Germany) which can test 10 different parameters including protein, blood, specific gravity, urobilinogen, leucocyte esterase, PH, nitrite, glucose, ketone, and bilirubin. All instructions regarding the handling of the reagent strip and performance of the dipstick will be strictly adhered to as stipulated by the manufacturer. The strip will be dipped briefly in the urine sample in the universal container making sure the test areas are fully immersed. Excess urine will be removed from the strip by running the edge of the strip along the rim of the urine container. The reaction will be read in good light between 30 and 60 seconds as stated by the manufacturer. The reaction will be compared by holding the strip close to the colour chart on the container label and the different parameters will be recorded. A urine dipstick that reads ≥ 2+ for blood or protein will be considered significant.87 Also the dipstick reading for other parameters especially specific gravity, PH, nitrite, and leukocyte esterase will be noted and documented.

Significant haematuria or proteinuria will be taken as haematuria or proteinuria of ≥2+ respectively while negative, trace, and 1+ will be taken as insignificant proteinuria or haematuria.

Hyposthenuria will be considered as urine specific gravity of less than 1.010 on dipstick urinalysis.

Calculation of eGFR

Cr estimated GFR will be calculated for each subject using the original Schwartz formula, taking into consideration the different values for the constant ‘K’ for each age group and gender: 0.55 for children aged 1-13 years and adolescent females above 13 years. ‘K’ is 0.70 for adolescent males above 13years.

GFR (mls/min/1.73m2)92 = K x height (cm) / SCr (mg/dl)

AKI will be defined using the Acute Kidney Injury Network (AKIN) classification as an increase in SCr by ≥0.3mg/dl (≥26.5µmol/l) within 48hours or 50% increase in serum creatinine .

Samples for serum creatinine estimation will be collected at presentation and repeated at 48hours of admission.

Renal dysfunction will be defined as hyperfiltration, renal insufficiency, AKI, significant proteinuria or haematuria, and hyposthenuria. Hyperfiltration is defined as GFR greater than 140mls/min/1.73m2 and renal insufficiency is defined as GFR less than 90 ml/min/1.73 m2.

All study participants with renal dysfunction will be counseled and consult will be sent to the paediatric nephrology unit for further evaluation and co-management with the haematology unit.

Data Analysis

The data obtained from the study will be entered into, cleaned, and analyzed using the Statistical Package for Social Sciences (SPSS) version 25.0.89 Results will be presented in tables and charts. Numerical variables will be summarized using mean and standard deviation, while categorical variables will be summarized using percentages. Chi-square will be used to determine the association between the presence of renal dysfunction and other categorical variables. Multivariable binary logistic regression analysis will be used to determine the predictors of renal dysfunction among the children. The level of significance will be set at 0,05.

Conclusion

this comprehensive study conducted at the Emergency Paediatric Unit of the National Hospital Abuja aims to shed light on the prevalence and predictors of renal dysfunction among children aged 2-15 years with sickle cell disease experiencing crises. Utilizing a cross-sectional descriptive study design, the research meticulously collected data from a carefully selected study population, adhering to stringent inclusion and exclusion criteria to ensure the reliability of the findings. The methodology encompassed a thorough data collection process, involving clinical examinations and laboratory assays to accurately assess the renal function and identify any dysfunction.

The calculated sample size of 114, factoring in an attrition rate, underscores the study's commitment to achieving statistically significant results. By employing the Statistical Package for Social Sciences (SPSS) for data analysis, the study promises to contribute valuable insights into the renal health of children with sickle cell disease in the FCT, potentially informing future healthcare practices and policies. The anticipation of identifying significant predictors of renal dysfunction underscores the study's potential impact on improving early diagnosis, treatment, and management strategies for this vulnerable population, thereby enhancing their quality of life and health outcomes.