Type 2 diabetes mellitus in sub-Saharan Africa: challenges and opportunities

Epidemiology of fractures in adults of African ancestry with diabetes mellitus: A systematic review and meta-analysis

Diabetes mellitus (DM) is associated with increased fracture risk in White adults. However, the impact of DM on fractures in Black adults is unknown. This systematic review and meta-analysis investigated the association between DM and fractures in adults of African ancestry. MEDLINE, Scopus, CINAHL and Embase databases were searched from their inception up to November 2023 for all studies in the English language investigating the epidemiology of fractures (prevalence, incidence, or risk) in Black men and women (age ≥ 18 years) with type 1 or type 2 DM. Effect sizes for prevalence of previous fractures (%) and incident fracture risk (hazard ratio [HR]) were calculated using a random-effects model on Stata (version 18.0). There were 13 eligible studies, of which 12 were conducted in Black adults from the United States, while one was conducted in adults of West African ancestry from Trinidad and Tobago. We found no fracture data in Black adults with DM living in Africa. Five studies were included in a meta-analysis of incident fracture risk, reporting data from 2926 Black and 6531 White adults with DM. There was increased risk of fractures in Black adults with DM compared to non-DM (HR = 1.65; 95 % confidence interval [CI]: 1.14, 2.39). The risk of fractures was also higher in White adults with DM compared to non-DM (HR = 1.31; 95 % CI: 1.06, 1.61) among these studies. Five studies were included in a meta-analysis of fracture prevalence, of which three also reported fracture prevalence in White adults. There were 175 previous fractures among 993 Black adults with DM and 384 previous fractures among 1467 White adults with DM, with a pooled prevalence of 17.5 % (95 % CI: 15.4, 19.6) and 25.8 % (95 % CI: 4.8, 46.8), respectively. Our results indicate a high burden of fractures in Black adults with DM.

Epidemiology of Non-Traumatic Lower Extremities Amputations in West Africa: Nationwide Data from Togo

OBJECTIVE

Non-traumatic lower limb amputation (NT-LLA) has consequences at individual and public health levels. Population based studies in sub-Saharan Africa are scarce and often related to single centre series. This study aimed to estimate the incidence of NT-LLA (minor and major) and to describe epidemiological, clinical, and prognostic aspects in Togo.

METHODS

This was a population based observational study conducted among all patients who underwent NT-LLA. Traumatic amputations were excluded. Sociodemographic, clinical, and work up data were collected from clinical files in any Togolese health centre from 1 January 2016 to 31 December 2021. Incidence rates were adjusted for age.

RESULTS

Over the six year period, 352 patients (59% males) underwent NT-LLA (mean ± standard deviation age 60 ± 15.7 years). The average age adjusted incidence rate of NT-LLA was 8.5 per million/year (95% confidence interval [CI] 7.6 - 9.4). Men were 1.7 times more likely to undergo a NT-LLA than women. The relative risk of NT-LLA was 48 times higher in patients with diabetes than in patients without diabetes. Around 61.0% of the NT-LLAs occurred within the 50 - 74 age group and 54.3% had diabetes mellitus. Among amputees, 54.5% had a diagnosis of peripheral artery disease (PAD) and 52.8% had diabetic ulcers, with co-existence of several factors. Less than 5% of participants had a history of smoking tobacco. Average length of hospital stay was 12 days. The in hospital mortality rate was 8.8% (9.0% for major, 6.7% for minor amputations). Only 18.2% had duplex ultrasound performed and 1.7% angiography prior to amputation. No patient underwent vascular intervention prior to amputation.

CONCLUSION

This is the first study to report nationwide and contemporary epidemiological data on NT-LLAs in West Africa, highlighting several specificities. Large scale interventions are needed to ameliorate the care of diabetes and PAD and improve facilities for optimal management of patients at risk of amputation in Africa.

Association between area under the C-peptide curve during an oral glucose tolerance test and diabetic retinopathy in type 2 diabetes patients

AIMS/INTRODUCTION

To evaluate the relative contributions of the area under the C-peptide curve (AUCC ) in diabetic retinopathy (DR) during an oral glucose tolerance test and C-peptide release test in patients with type 2 diabetes.

MATERIALS AND METHODS

We retrospectively analyzed the data of 969 patients. Their general characteristics were retrieved. A series of parameters for assessing pancreatic β-cells function, such as the AUCC for six time periods: 0-60 min (AUCC0-60 ), 0-120 min (AUCC0-120 ), 0-180 min (AUCC0-180 ), 60-120 min (AUCC60-120 ), 60-180 min (AUCC60-180 ) and 120-180 min (AUCC120-180 ); the area under the glucose-time curve for six time periods: 0-60 min (AUCG0-60 ), 0-120 min (AUCG0-120 ), 0-180 min (AUCG0-180 ), 60-120 min (AUCG60-120 ), 60-180 min (AUCG60-180 ) and 120-180 min (AUCG120-180 ) and their related indexes, were calculated through 0-180 min oral glucose tolerance test and C-peptide release test. We used univariate analysis to examine the potential factors affecting DR. Spearman's correlation was used to analyze the correlation between AUCC -related indexes and DR. The logistic regression model was used to investigate AUCC and its related indexes' contribution to incidence DR. A smooth curve fitting model was used to determine the correlation, non-linear relationship, and threshold effect between AUCC and DR.

RESULTS

Of the 969 patients with type 2 diabetes, 469 (48.40%) and 500 (51.60%) were classified as the DR group and non-DR group. Compared with the non-DR group, the DR patients had lower AUCC and AUCC /AUCG . Spearman's correlation analysis showed that AUCC -related indexes were all negatively correlated with DR. The logistic regression analysis determined that there were associations between AUCC and DR in the adjusted models. The odds ratio values of AUCC0-60 , AUCC0-120 , AUCC0-180 , AUCC0-60 /AUCG0-60 , AUCC0-120 /AUCG0-120 , AUCC0-180 /AUCG0-180 , AUCC60-120 , AUCC60-180 , AUCC120-180 , AUCC60-120 /AUCG60-120 , AUCC60-180 /AUCG60-180 and AUCC120-180 /AUCG120-180 were 0.817 (0.750, 0.890), 0.925 (0.895, 0.955), 0.951 (0.932, 0.970), 0.143 (0.060, 0.340), 0.194 (0.093, 0.406), 0.223 (0.116, 0.427), 0.886 (0.842, 0.933), 0.939 (0.915, 0.963), 0.887 (0.846, 0.930), 0.253 (0.133, 0.479), 0.282 (0.160, 0.497) and 0.355 (0.220, 0.573), respectively. AUCC showed a non-linear relationship with DR, with an inflection point. The inflection points of AUCC180 /AUCG180 , AUCC60-120 , AUCC60-180 , AUCC120-180 , AUCC60-120 /AUCG60-120 , AUCC60-180 /AUCG60-180 , AUCC120-180 /AUCG120-180 and DR were 17.51, 0.542, 6.6, 15.7, 8.23, 0.534, 0.593 and 0.808 (P < 0.0001). When the indexes related to the AUCC were less than the inflection point value, they were significantly negatively associated with DR.

CONCLUSIONS

The indexes related to the AUCC for six time periods during an oral glucose tolerance test and C-peptide release test was closely associated with the incidence to DR in patients with type 2 diabetes. AUCC has the added advantage of being a cheap and convenient risk assessment over traditional ophthalmic screening.

Prevalence of obesity, hypertension and diabetes among people living with HIV in South Africa: a systematic review and meta-analysis

BACKGROUND

HIV has become a manageable chronic condition due to the success and scale-up of antiretroviral therapy (ART). Globally, South Africa has the highest number of people living with HIV (PLHIV) and research evidence indicates that countries with the highest burden of PLHIV have a substantial burden of obesity, hypertension (HPT) and type 2 diabetes (T2D). We sought to summarize the burden of these three common NCDs among PLHIV in South Africa.

METHODS

In this systematic review, multiple databases were searched for articles reporting on the prevalence of obesity, HPT, and T2D among PLHIV in South Africa published since journal inception until March 2022. A meta-analysis was conducted using random-effects models to obtain pooled prevalence estimates of the three NCDs. Heterogeneity was assessed using X2 test on Cochran's Q statistic.

RESULTS

We included 32 studies, with 19, 22 and 18 studies reporting the prevalence of obesity, HPT, and T2D among PLHIV, respectively. The overall prevalence of obesity, HPT, and T2D was 23.2% [95% CI 17.6; 29.9], 25.5% [95% CI 15.6; 38.7], and 6.1% [95% CI 3.8; 9.7] respectively. The prevalence of obesity was significantly higher among women (P = 0.034) compared to men, however the prevalence of HPT and T2D did not differ by sex. The prevalence of each of the three NCDs did not differ significantly between rural, urban, and peri-urban areas. The prevalence of obesity and T2D was higher in studies conducted between 2013 and 2022 compared to studies conducted between 2000 and 2012, while the prevalence of HPT was higher between 2000 and 2012 compared to between 2013 and 2022.

CONCLUSIONS

These findings suggest that South Africa is experiencing a syndemic of NCDs among people PLHIV highlighting the need to increase cost-effective interventions and management strategies that involve integrated HIV and NCD care in the South African setting.

Type 2 diabetes in the Democratic Republic of Congo: an urgent need for a management framework

Glycaemic control is of one the main goals for managing type 2 diabetes. In sub-Saharan Africa and the Democratic Republic of the Congo, studies have reported alarming poor control rates. Patients with poor glycaemic control are exposed to complications leading to high cost of care and deteriorated quality of life. In recent studies by our group, we have demonstrated that poor glycaemic control is high and driven by proximal (individual) and distal (structural) factors in Kinshasa, Democratic Republic of the Congo. Financial constraints impacted many aspects of care at multiple levels from the Government to persons living with diabetes. Financial constraints prevented good preparation, organization and access to diabetes care. Difficulties in implementing lifestyle changes, lack of health literacy and limited healthcare support were also contributing to poor glycaemic control. Through a Delphi study, a group of experts reached a consensus on five potential strategies for improving glycaemic control in the Democratic Republic of Congo as follows: changing the healthcare system for better diabetes care extended to other noncommunicable diseases, ensuring consistent financing of the healthcare, augmenting the awareness of diabetes among the general population and the persons living with diabetes, easing the adoption of lifestyle modifications and reducing the burden of undiagnosed diabetes. This paper reflects on the urgent need for an improved management framework for diabetes care in the Democratic Republic of the Congo. Specifically, the Government needs to increase the investment in the prevention and treatment of noncommunicable diseases including diabetes.

Global variation in diabetes diagnosis and prevalence based on fasting glucose and hemoglobin A1c

Fasting plasma glucose (FPG) and hemoglobin A1c (HbA1c) are both used to diagnose diabetes, but these measurements can identify different people as having diabetes. We used data from 117 population-based studies and quantified, in different world regions, the prevalence of diagnosed diabetes, and whether those who were previously undiagnosed and detected as having diabetes in survey screening, had elevated FPG, HbA1c or both. We developed prediction equations for estimating the probability that a person without previously diagnosed diabetes, and at a specific level of FPG, had elevated HbA1c, and vice versa. The age-standardized proportion of diabetes that was previously undiagnosed and detected in survey screening ranged from 30% in the high-income western region to 66% in south Asia. Among those with screen-detected diabetes with either test, the age-standardized proportion who had elevated levels of both FPG and HbA1c was 29-39% across regions; the remainder had discordant elevation of FPG or HbA1c. In most low- and middle-income regions, isolated elevated HbA1c was more common than isolated elevated FPG. In these regions, the use of FPG alone may delay diabetes diagnosis and underestimate diabetes prevalence. Our prediction equations help allocate finite resources for measuring HbA1c to reduce the global shortfall in diabetes diagnosis and surveillance.

Tuberculosis and COVID-19 in the elderly: factors driving a higher burden of disease

Mycobacterium tuberculosis (M.tb) and SARS-CoV-2 are both infections that can lead to severe disease in the lower lung. However, these two infections are caused by very different pathogens (Mycobacterium vs. virus), they have different mechanisms of pathogenesis and immune response, and differ in how long the infection lasts. Despite the differences, SARS-CoV-2 and M.tb share a common feature, which is also frequently observed in other respiratory infections: the burden of disease in the elderly is greater. Here, we discuss possible reasons for the higher burden in older adults, including the effect of co-morbidities, deterioration of the lung environment, auto-immunity, and a reduced antibody response. While the answer is likely to be multifactorial, understanding the main drivers across different infections may allow us to design broader interventions that increase the health-span of older people.

Implementing Innovative Approaches to Improve Health Care Delivery Systems for Integrating Communicable and Non-Communicable Diseases Using Tuberculosis and Diabetes as a Model in Tanzania

Background: Many evidence-based health interventions, particularly in low-income settings, have failed to deliver the expected impact. We designed an Adaptive Diseases Control Expert Programme in Tanzania (ADEPT) to address systemic challenges in health care delivery and examined the feasibility, acceptability and effectiveness of the model using tuberculosis (TB) and diabetes mellitus (DM) as a prototype. Methods: This was an effectiveness-implementation hybrid type-3 design that was implemented in Dar es Salaam, Iringa and Kilimanjaro regions. The strategy included a stepwise training approach with web-based platforms adapting the Gibbs' reflective cycle. Health facilities with TB services were supplemented with DM diagnostics, including glycated haemoglobin A1c (HbA1c). The clinical audit was deployed as a measure of fidelity. Retrospective and cross-sectional designs were used to assess the fidelity, acceptability and feasibility of the model. Results: From 2019-2021, the clinical audit showed that ADEPT intervention health facilities more often identified median 8 (IQR 6-19) individuals with dual TB and DM, compared with control health facilities, median of 1 (IQR 0-3) (p = 0.02). Likewise, the clinical utility of HbA1c on intervention sites was 63% (IQR:35-75%) in TB/DM individuals compared to none in the control sites at all levels, whereas other components of the standard of clinical management of patients with dual TB and DM did not significantly differ. The health facilities showed no difference in screening for additional comorbidities such as hypertension and malnutrition. The stepwise training enrolled a total of 46 nurse officers and medical doctors/specialists for web-based training and 40 (87%) attended the workshop. Thirty-one (67%), 18 nurse officers and 13 medical doctors/specialists, implemented the second step of training others and yielded a total of 519 additional front-line health care workers trained: 371 nurses and 148 clinicians. Overall, the ADEPT model was scored as feasible by metrics applied to both front-line health care providers and health facilities. Conclusions: It was feasible to use a stepwise training and clinical audit to support the integration of TB and DM management and it was largely acceptable and effective in differing regions within Tanzania. When adapted in the Tanzania health system context, the model will likely improve quality of services.

Quality of care of patients with type 2 diabetes mellitus at a public sector district hospital

BACKGROUND

 Globally, diabetes mellitus (DM) remains one of the leading causes of mortality, with approximately 2 million deaths in 2019, the condition also contributes significantly to adverse health conditions and costs. The study aimed to describe the quality of care (QOC) rendered to patients with type 2 DM (T2DM) seeking care at Wentworth Hospital (WWH), a district hospital in KwaZulu-Natal province, South Africa.

METHODS

 A descriptive cross-sectional design was used, and all patients living with T2DM on treatment who had accessed care for at least 1 year were included. Data were collected through structured exit interviews, and their clinical data were extracted from their medical records. Their knowledge, attitudes and practices were assessed using a 5-point Likert scale.

RESULTS

 The mean age (standard deviation [s.d.]) was 59 (13.0) years and most (65.3%) were female, of African (30.0%) and Indian (38.6%) descent, with two-thirds (69.4%) obtaining a secondary school education. Their mean glycated haemoglobin (HbA1c) (s.d.) was 8.6 (2.4%). Over 82% had one or more comorbidity, while 30% had at least one DM-related complication. Generally, participants were pleased with the care received, but their knowledge and practices related to their T2DM was suboptimal.

CONCLUSION

 This study indicates that the QOC was suboptimal due to poor efficacy indicators, poor knowledge and lack of adequate lifestyle measures, despite the frequency of medical practitioner reviews.Contributions: This study identified gaps in QOC and will aid South African public sector policy-makers in devising quality improvement initiatives.

Type 2 diabetes prevalence, awareness, and risk factors in rural Mali: a cross-sectional study

Diabetes is currently a crisis in sub-Saharan West Africa (SSWA) with dramatic implications for public health and national budgets prioritizing infectious diseases. There is limited recent literature about the prevalence, awareness, and risk factors for type 2 diabetes (T2D) in rural parts of SSWA. This study characterized T2D prevalence and risk factors for the rural Malian community of Nièna, which is situated in Mali's second-largest province of Sikasso. Between December 2020 and July 2021, a cross-sectional study of 412 participants was conducted in the Nièna community using clinical questionnaires and rapid diagnostic tests. Among 412 participants, there were 143 (34.7%) and 269 (65.3%) males and females, respectively. The overall prevalence of T2D in Nièna was 7.5% (31/412), and prevalence rates were 8.6% (23/269) and 5.6% (8/143) for females and males, respectively. Age, family history of diabetes, hypertension, waist circumference, and fetal macrosomia were significantly associated with T2D (p = 0.007, p < 0.001, p = 0.003, p = 0.013, and p < 0.001, respectively). Notably, 61.3% (19/31) of T2D subjects were unaware of their diabetic status before the study. Field surveys have considerable utility in driving T2D awareness in rural African settings.

Change in glycaemic control with structured diabetes self-management education in urban low-resource settings: multicentre randomised trial of effectiveness

BACKGROUND

In high-resource settings, structured diabetes self-management education is associated with improved outcomes but the evidence from low-resource settings is limited and inconclusive.

AIM

To compare, structured diabetes self-management education to usual care, in adults with type 2 diabetes, in low-resource settings.

RESEARCH DESIGN AND METHODS

DESIGN

Single-blind randomised parallel comparator controlled multi-centre trial. Adults (> 18 years) with type 2 diabetes from two hospitals in urban Ghana were randomised 1:1 to usual care only, or usual care plus a structured diabetes self-management education program. Randomisation codes were computer-generated, and allotment concealed in opaque numbered envelopes. The intervention effect was assessed with linear mixed models.

MAIN OUTCOME

Change in HbA1c after 3-month follow-up. Primary analysis involved all participants.

CLINICALTRIAL

gov identifier:NCT04780425, retrospectively registered on 03/03/2021.

RESULTS

Recruitment: 22^nd until 29^th January 2021. We randomised 206 participants (69% female, median age 58 years [IQR: 49-64], baseline HbA1c median 64 mmol/mol [IQR: 45-88 mmol/mol],7.9%[IQR: 6.4-10.2]). Primary outcome data was available for 79 and 80 participants in the intervention and control groups, respectively. Reasons for loss to follow-up were death (n = 1), stroke(n = 1) and unreachable or unavailable (n = 47). A reduction in HbA1c was found in both groups; -9 mmol/mol [95% CI: -13 to -5 mmol/mol], -0·9% [95% CI: -1·2% to -0·51%] in the intervention group and -3 mmol/mol [95% CI -6 to 1 mmol/mol], -0·3% [95% CI: -0·6% to 0.0%] in the control group. The intervention effect was 1 mmol/mol [95%CI:-5 TO 8 p = 0.726]; 0.1% [95% CI: -0.5, 0.7], p = 0·724], adjusted for site, age, and duration of diabetes. No significant harms were observed.

CONCLUSION

In low-resource settings, diabetes self-management education might not be associated with glycaemic control. Clinician's expectations from diabetes self-management education must therefore be guarded.

Health Care Resource Utilization in Adults Living With Type 1 Diabetes Mellitus in the South African Public Health Sector: Protocol for a 1-Year Retrospective Analysis With a 5-, 10-, and 25-Year Projection

BACKGROUND

Type 1 diabetes mellitus (T1DM) is less common than type 2 diabetes mellitus but is increasing in frequency in South Africa. It tends to affect younger individuals, and upon diagnosis, exogenous insulin is essential for survival. In South Africa, the health care system is divided into private and public health care systems. The private system is well resourced, whereas the public sector, which treats more than 80% of the population, has minimal resources. There are currently no studies in South Africa, and Africa at large, that have evaluated the immediate and long-term costs of managing people living with T1DM in the public sector.

OBJECTIVE

The primary objective was to quantify the cost of health care resource utilization over a 12-month period in patients with controlled and uncontrolled T1DM in the public health care sector. In addition, we will project costs for 5, 10, and 25 years and determine if there are cost differences in managing subsets of patients who achieve glycemic control (hemoglobin A_1c [HbA_1c] <7%) and those who do not.

METHODS

The study was performed in accordance with Good Epidemiological Practice. Ethical clearance and institutional permissions were acquired. Clinical data were collected from 2 tertiary hospitals in South Africa. Patients with T1DM, who provided written informed consent, and who satisfied the inclusion criteria were enrolled in the study. Data collection included demographic and clinical characteristics, acute and chronic complications, hospital admissions, and so on. We plan to perform a cost-effectiveness analysis to quantify the costs of health care utilization in the preceding 12 months. In addition, we will estimate projected costs over the next 10 years, assuming that study participants maintain their current HbA_1c level. The cost-effectiveness analysis will be modeled using the IQVIA CORE Diabetes Model. The primary outcome measures are incremental quality-adjusted life years, incremental costs, incremental cost-effectiveness ratios, and incremental life years.

RESULTS

Ethical clearance and institutional approval were obtained (reference number 200407). Enrollment began on February 9, 2021, and was completed on August 24, 2021, with 224 participants. A database lock was performed on October 29, 2021. The statistical analysis and clinical study report were completed in January 2022.

CONCLUSIONS

At present, there are no data assessing the short- and long-term costs of managing patients with T1DM in the South African public sector. It is hoped that the findings of this study will help policy makers optimally use limited resources to reduce morbidity and mortality in people living with T1DM.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR1-10.2196/44308.

Advanced Drug Delivery System for Management of Chronic Diabetes Wound Healing

The diabetic wound is excessively vulnerable to infection because the diabetic wound suggests delayed and incomplete healing techniques. Presently, wounds and ulcers related to diabetes have additionally increased the medical burden. A diabetic wound can impair mobility, lead to amputations, or even death. In recent times, advanced drug delivery systems have emerged as promising approaches for enhancing the efficacy of wound healing treatments in diabetic patients. This review aims to provide an overview of the current advancements in drug delivery systems in managing chronic diabetic wound healing. This review begins by discussing the pathophysiological features of diabetic wounds, including impaired angiogenesis, elevated reactive oxygen species, and compromised immune response. These factors contribute to delayed wound healing and increased susceptibility to infection. The importance of early intervention and effective wound management strategies is emphasized. Various types of advanced drug delivery systems are then explored, including nanoparticles, hydrogels, transferosomes, liposomes, niosomes, dendrimers, and nanosuspension with incorporated bioactive agents and biological macromolecules are also utilized for chronic diabetes wound management. These systems offer advantages such as sustained release of therapeutic agents, improved targeting and penetration, and enhanced wound closure. Additionally, the review highlights the potential of novel approaches such as antibiotics, minerals, vitamins, growth factors gene therapy, and stem cell-based therapy in diabetic wound healing. The outcome of advanced drug delivery systems holds immense potential in managing chronic diabetic wound healing. They offer innovative approaches for delivering therapeutic agents, improving wound closure, and addressing the specific pathophysiological characteristics of diabetic wounds.

Detecting High-Risk Factors and Early Diagnosis of Diabetes Using Machine Learning Methods

Diabetes is a chronic disease that can cause several forms of chronic damage to the human body, including heart problems, kidney failure, depression, eye damage, and nerve damage. There are several risk factors involved in causing this disease, with some of the most common being obesity, age, insulin resistance, and hypertension. Therefore, early detection of these risk factors is vital in helping patients reverse diabetes from the early stage to live healthy lives. Machine learning (ML) is a useful tool that can easily detect diabetes from several risk factors and, based on the findings, provide a decision-based model that can help in diagnosing the disease. This study aims to detect the risk factors of diabetes using ML methods and to provide a decision support system for medical practitioners that can help them in diagnosing diabetes. Moreover, besides various other preprocessing steps, this study has used the synthetic minority over-sampling technique integrated with the edited nearest neighbor (SMOTE-ENN) method for balancing the BRFSS dataset. The SMOTE-ENN is a more powerful method than the individual SMOTE method. Several ML methods were applied to the processed BRFSS dataset and built prediction models for detecting the risk factors that can help in diagnosing diabetes patients in the early stage. The prediction models were evaluated using various measures that show the high performance of the models. The experimental results show the reliability of the proposed models, demonstrating that k-nearest neighbor (KNN) outperformed other methods with an accuracy of 98.38%, sensitivity, specificity, and ROC/AUC score of 98%. Moreover, compared with the existing state-of-the-art methods, the results confirm the efficacy of the proposed models in terms of accuracy and other evaluation measures. The use of SMOTE-ENN is more beneficial for balancing the dataset to build more accurate prediction models. This was the main reason it was possible to achieve models more accurate than the existing ones.