Strengthening Health Systems for Quality Health Care: A Study of Misdiagnosis among Hospitalised Patients in General Hospitals in Uganda

Abstract

Background: Achieving high-quality healthcare that is safe, effective, efficient, patient-centered, timely, and equitable necessitates a systemic approach. Not many countries, particularly in sub Saharan Africa, including Uganda optimally achieve this. Poor quality of healthcare characterized by inappropriate treatments, medical errors and poor outcomes has often resulted from patient misdiagnosis. An estimated 5-15% of hospitalized patients globally, are misdiagnosed posing a significant barrier to quality and safe healthcare. Misdiagnosis is particularly concerning in such resource-constrained healthcare systems as that of Uganda, where it can hinder progress towards attaining universal healthcare coverage. This study, therefore, was conducted with a general objective of establishing the prevalence of patient misdiagnosis and factors contributing to misdiagnosis among hospitalized patients in general hospitals in Uganda. The ultimate goal is to provide information that can contribute to strengthening healthcare systems for quality healthcare delivery. Methods: The study used an explanatory mixed-method cross-sectional research design, collecting both quantitative and qualitative data, to examine misdiagnosis in Kiboga, Nakaseke, Gombe, Kayunga, and Mityana general hospitals. Stratified and simple random sampling techniques were used to sample the medical records, ensuring representativeness. Stratified random sampling considered the hospital of admission, the patient's age, and their gender to ensure a balanced representation. As part of this method, proportional allocation was used for comparable sample sizes, including children aged 0–9 and those over 10 years old. An allocation of 50% between males and females promoted balanced comparisons across genders, an essential factor in research that evaluates conditions or treatments. Records of 2,431 patients admitted between July 1, 2019 and June 30, 2020 were analyzed to determine the proportion of misdiagnosed patients. The original diagnosis assigned by the clinician or diagnostician was compared to the diagnosis confirmed by a medical officer or physician on the ward. Misdiagnosis was categorized into four levels of Class I; Class II, Class III, and Class IV guided by the International Classification of Diseases version eleven (ICD-11). The Pareto principle was used to identify the most common misdiagnosed diseases and major diagnostic groupings. Logistic regression was used to analyze the factors associated with misdiagnosis, with significant variables with a p-value of ≤0.05 and their adjusted odds ratios considered as independently associated factors. In-depth interviews were conducted with eight clinical officers and seven medical officers to provide insights into the phenomenon of misdiagnosis in the hospitals. The qualitative analysis employed a deductive thematic approach, with data analyzed manually and the identified themes systematically aligned with the health system building blocks, as well as the structure, processes, and outcomes of the Safer Diagnostic framework. Results: The study observed that 9.2% (223 out of 2,431; 95% CI of 8.1-10.3%) of patients were misdiagnosed, with 70.9% (158 out of 223) of these misdiagnosed patients classified as Class I, indicating that the wrong diagnosis and the correct diagnosis belonged to different major diagnostic groupings. Infectious or parasitic diseases (32%), digestive system diseases (12%), circulatory system diseases (11%), endocrine, nutritional, or metabolic disorders (9%), genitourinary system diseases (7%), respiratory system diseases (7%), and blood and blood forming organ diseases (5%), were the seven most misdiagnosed major diagnostic groupings. Peptic ulcer disease, severe malaria, hypertension, gastroenteritis, and pneumonia were among the most misdiagnosed diseases. Non-communicable diseases accounted for most misdiagnosed conditions (82.9%). The multivariable logistic regression analysis revealed that variables associated with misdiagnosis included; being admitted to Nakaseke hospital (1.95 times more likely; 95% CI =1.17-3.25, P=.01, being admitted at night (3 times more likely; 95% CI=1.81-5.02, P<.001), being male (1.89 times more likely; 95% CI=1.35-2.64, P<.001), being in the age group of 10 to 50 years and above (aOR 2.3; 95% CI=2.3-9.25, P<.001; aOR 8.15; 95% CI=4.18-15.89, P< .001; aOR 8.12; 95% CI=3.99- 16.54, P<.001; aOR 7.88; 95% CI=3.71-16.73, P<.001; aOR 12.14; 95% CI=6.41-23.01, P<.001). Other variables associated with misdiagnosis included patients with multimorbidity (aOR 4.71; 95% CI=1.91-11.65, P<.001) and patients treated for uncommon diseases of admission to the hospital (aOR 2.57 95% CI=1.28-5.18, P<0.001). On the other hand, patients without underlying diseases were 37% less likely to be misdiagnosed (95% CI=0.43-0.91, P=0.015), and patients who were not referred to the hospital were 49% less likely to be misdiagnosed compared to those who were referred (95% CI=0.31-0.86, P=0.011). The phenomenon of patient misdiagnosis, classification and associated factors is complex and was explained by the diagnosticians, analyzed and presented through the lenses of Safer Diagnostic Framework and health system building blocks. The issue of advancing age and its association with misdiagnosis was explained to be due to complex or uncommon diseases in such age groups, which clinicians are unfamiliar with compared to less complex and easy to diagnose diseases in children. On the other hand, the high likelihood of misdiagnosis among male patients was due to their poor healthcare-seeking behaviours, especially late reporting to the hospital when severely sick complicating the diagnostic process, and by their unfamiliarity with healthcare systems. Human resources for health issues contributing to misdiagnosis included inadequate healthcare workforce numbersleading to work overload, shortfalls in training, and healthcare workers fatigue, especially at night. Besides, other factors contributing to misdiagnosis cited included inadequate infrastructure, particularly laboratory and radiological examinations, and other diagnostic tools and technology. Several organizational or hospital factors, including suboptimal nighttime service arrangements, and the absence of a quality of care and safety culture, underscored the role of leadership and governance play in misdiagnosis. Additionally, inadequate service delivery such as lack of investigations and clinical support hamper appropriate diagnosis at night. Challenges in the patient referral process from primary healthcare facilities to hospitals, including patients being referred with incomplete referral notes, could contribute to the high risk of misdiagnosing referred patients, among other factors. Conclusion: Misdiagnosis in the general hospitals studied remains a significant quality of healthcare and safety problem that has a complex and of multifactorial etiology; mostly attributable to seven major diagnostic groupings and 19 diseases. While the prevalence fell within global estimated ranges, it was concerning that most misdiagnoses were classified as Class I. The result of the discovery of this misdiagnosis was that significant treatment changes were needed. In the absence of any changes in patient management, especially if the misdiagnosis was not recognized, a poor outcome would have resulted. Predispositions within the leadership and governance, service delivery, medicines and health technologies, and human resources for health underpin the observed misdiagnosis occurrence in this study. To enhance patient diagnosis and treatment outcomes, health services planners, managers and implementers should prioritize this health system building blocks, with particular focus on the 19 most commonly misdiagnosed conditions and major diagnostic groups.