Magnitude and Factors Associated With Drop Out of Care Among Children in Long Term Care at Kenyatta National Hospital Diabetic Clinic Between 2012- 2022

Abstract

Background: Diabetes Mellitus being a chronic disease, DM patients require close monitoring and long term follow up for optimal glycemic control. Suboptimal glycemic control leads to multisystem complications. DM can be divided in two main groups: type 1 and type 2. Other types are neonatal diabetes and maturity onset diabetes of the young. Type 1 is more common in children with a prevalence of 4.5% of the world population below 14 years old. The prevalence of LTFU in developed countries ranges between 5- 46% and in Kenya it is estimated at 31% in adults. There is limited data on prevalence of lost to follow up in paediatric patients in general and more so in diabetic patients as well as associated factors. Objectives: Primary: To determine the incidence and the factors associated with retention versus drop out of care among children<16 years with Diabetes mellitus who received care at Kenyatta National Hospital endocrinology clinic during a 10 year period-2012 to 2022. Secondary: To determine the reasons behind dropping out of care from Kenyatta National Hospital among those children who are traceable by telephone. To determine the perceptions of health care workers on facilitators and barriers to retention in care for diabetic children in care at Kenyatta National Hospital Study design: We used a multi-design approach in which we conducted 1) a retrospective cohort study to determine the incidence and factors associated with drop out of care by abstraction of medical records. 2) A descriptive cross sectional design was used to determine the reasons for drop out of care by semi structured interview of parents of LTFU children who were successfully reached through phone call. And we conducted FGD involving healthcare workers for facilitators and barriers to retention in care for diabetic children Method: We included all diabetics patients aged< 16 years followed up at the diabetic clinic in Kenyatta National Hospital between 2012- 2022 excluding those whose files were missing, those with one visit and those who do not give informed consent. For objective 1 and 2, we obtained sociodemographic and clinical information from the patient file. For objective 3, a phone administered questionnaire was used to interview the caregivers of children who had dropped out of care successfully traced. And for objective 4, we conducted focus group discussion for health care workers in diabetic clinic on barriers to retention in care. Ethical approval was obtained from both KNH ERC and the department of paediatrics. We calculated

the incidence of LTFU over the 10 years period. We performed logistic regression to identify factors associated with LTFU. Data from focus group discussion was manually summarized into thematic areas. Results: From January 2012 to December 2022, 273 diabetic children were followed up at KNH endocrinology clinic, with 146 (53.5%) % being female. The mean age was 7.2(SD 4.3) years. The overall follow up time of the whole cohort was 654.5 person years with a median of 6.9 (95% CI: 4.6-9.3; IQR 1.0-9.9) years. By December 2022, 146 children (53.4%) remained in care. Out of the 127 children presumed as LTFU, we successfully traced 38 (29.9%). Two (1.5%) children were transferred out, 30(23.6%) died and the remainder 95 (74.8%) were classified as true LTFU giving a proportion of true LTFU of 34.8% (95%CI 29.4%-40.6%) from the total sample size and an incidence rate of 14.5(95% CI 12.0%-17.4%) per 100 person years. Independent factors that were predictive of LTFU at univariate analysis included lack of medical insurance (OR=6.04, 95% CI: 3.47-10.53, p< 0.001), and while presence of complications was associated with retention in care (OR=0.57, 95% CI: 0.34-0.95, p=0.032). On subjecting the 2 significant factors on multivariate analysis, only lack of medical insurance (OR=5.79, 95% CI: 3.27-10.24, p<0.001) remained statistically significant associated with loss to follow up. From caregivers tracing, half of the caregivers cited financial constraint as major reason, followed by long distance and self transfer. From the focus group discussion, the factors perceived by healthcare workers were low socio-economic status, shortage of staff and poor tracing system. Conclusion: Over the 10 years period, the incidence of LTFU of diabetic children was 14.5 per 100 person years and overall, the factor statistically associated with drop out of care was lack of medical insurance and the themes emerging from both caregivers tracing and the FGD were low socio-economic status, long distance, shortage of staff and poor tracing system. Conclusively, based on quantitative and qualitative analysis in our study, LTFU among DM children at KNH is high and is associated with multiple factors. Several limitations were encountered during our study, additional studies including physical tracing are therefore needed to further investigate and understand in depth the factors associated with loss to follow up from caregivers/ patients