Factors associated with misdiagnosis of hospitalised patients: A mixed method study in five general hospitals of Central Uganda

doi:10.21203/rs.3.rs-2505054/v1

Download PDF

Research Article

Factors associated with misdiagnosis of hospitalised patients: A mixed method study in five general hospitals of Central Uganda

https://doi.org/10.21203/rs.3.rs-2505054/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Introduction: Inpatient misdiagnosis is a major public health problem in sub-Saharan Africa, the scope and causes of which are unknown. The purpose of this cross-sectional study, conducted in five hospitals in central Uganda, was to identify the factors associated with inpatient misdiagnosis in general hospitals in central Uganda.

Methodology: An explanatory mixed methods cross-sectional study was used. A retrospective review of 2,431 patient records was performed using explicit review methods to determine the extent of patient misdiagnosis and other variables thought to be related to patient misdiagnosis. Any discrepancy between the admission diagnosis made in the emergency room or outpatient clinic and the patient's discharge diagnosis made upon discharge was defined as a misdiagnosis. The diagnoses were classified using the World Health Organization ICD-11. Six clinical staff and six medical staff were interviewed using in-depth interviews (IDIs) in a phenomenological approach to obtain their explanatory account of factors associated with misdiagnosis. A logistic regression and a deductive thematic analysis were carried out on quantitative and qualitative data analyses. The quantitative and qualitative findings of the study were mixed in interpretation.

Results: Misdiagnosis was discovered in 223/2431 (9.2%) of the admitted patients' records. Misdiagnosis was associated with the following factors: a patient admitted to Nakaseke hospital [aOR = 1.95, 95% CI = 1.17–3.25, p = 0.01], admission at night [aOR = 3, 95% CI = 1.81–5.02, p0.01], male patient [aOR = 1.89, 95% CI = 1.35–2.64, p0.01], patient's age groups 10–19 [AOR = 2.3, 95% CI = 2.3-9. Misdiagnosis was also linked to multimorbidity (aOR = 4.71, 95% CI = 1.91–11.65, p0.01) and patients with rare diseases (aOR = 2.57, 95% CI = 1.28–5.18, p0.01). Patients with no underlying diseases [aOR = 0.63; 95% CI = 0.43–0.91, p = 0.015] and those who were not referred [aOR = 0.51; 95% CI = 0.31–0.86, p = 0.011] had a lower risk of misdiagnosis. The quantitative findings of significant (p 0.05) and non-significant patient, contextual, disease, and health system factors associated with misdiagnosis were validated by the qualitative findings.

Conclusion: To improve diagnostic accuracy, hospitals should reorganize patient admission processes, provide targeted training, create policies or guidelines targeting risk factors for misdiagnosis, and implement a diagnostic error prevention culture that addresses factors related to misdiagnosis in the respective hospitals oriented.

Over 5–15% of patients in hospitals around the world are misdiagnosed, leading to inadequate care and poor health outcomes (1). Unfortunately, research in low-income countries lacks sufficient evidence to help practitioners and policymakers fully understand the burden that misdiagnosis places on clinical practice and healthcare delivery (2, 3). While misdiagnosis is generally underreported in sub-Saharan Africa, for some diseases, such as HIV, it has been reported to range from 0.1–6.6% (4). The misdiagnosis of cancer in 34 patients initially diagnosed with tuberculosis in Malawi underscores the magnitude of the problem in sub-Saharan Africa (5). Another study in Uganda found that 30.6% of HIV-associated lymphoma patients were misdiagnosed with tuberculosis (6). In rural Uganda, up to 62% of febrile illnesses were diagnosed as malaria when malaria was not present (7). However, because most of these studies are disease-specific, it is difficult to determine the exact magnitude of the problem across the disease spectrum.

Diagnostic error is multifactorial in nature, with many levels of complexity and health system deficiencies affecting the incidence of misdiagnosis (8–10). Unfortunately, until now, misdiagnosis has been viewed from an individual perspective, with those on the frontline where misdiagnosis occurs being referred to as the sharp end to blame for any errors that occur (11). The majority of people are interested in who made a diagnostic error and what happened next, with less emphasis on the factors that may have contributed to the error (12).

Misdiagnosis-related patient harm, on the other hand, could be avoided more effectively if the factors associated with misdiagnosis were established using empirical evidence obtained from the departments, organizations, or health care systems where the misdiagnosis occurs (13). On this basis, the landmark Institute of Medicine (IOM) report "to err is human" advocated for identifying and correcting health-care system and contextual factors that may have contributed to diagnostic error (4). This study examines misdiagnosis in Ugandan hospitals from the perspectives of health systems, patients, diseases, and other contextual factors in order to inform processes aimed at improving patient diagnosis in low-income and Sub-Saharan African general hospitals and related healthcare systems.

Study design

This is an explanatory mixed-methods cross-sectional sequential study that included a review of patient charts admitted to the hospitals between July 1, 2019 and June 30, 2020, with the quantitative study being conducted first, followed by a qualitative study to explain the quantitative results. This is part of a doctoral research of the first author part of which has already been published (14).

Setting

It was carried out in the general hospitals of Mityana, Gombe, Nakaseke, Kiboga and Kayunga in central Uganda. In Uganda, a general hospital with a capacity of 150 beds serves as a district referral hospital for a population of over 500,000 people (15). It serves as a reference center for the lower health facilities of the district in which it is located, as well as providing inpatient and outpatient care for the residents of its catchment area. The central region was chosen for the study because it is more populous, has the most general hospitals, and has a diverse sociocultural environment, all of which could impact patient care

Quantitative Data

The sample size of the number of records to review was determined using the formula N = 10k/p based on the event per variable (EPV) for the binary logistic regression model to identify the factors associated with patient misdiagnosis. In this formula, p represents the smallest fraction of the desired outcome, while k represents the number of covariates (the number of independent variables) associated with the outcome. A given number of EPVs requires at least 10 cases per variable, resulting in 120 misdiagnosed cases (10 x 12)(16).

It has been reported that between 5% and 15% of patients are misdiagnosed, according to an experienced physician who changes the initial diagnosis (17). Therefore, 5% was chosen as the smallest possible outcome proportion (p). For 12 variables examined related to misdiagnosis, the required minimum number of patient records was 2,400, but 31 additional records were included to account for missing data and poor record quality. Records that met the inclusion criteria were selected from a pool of charts using simple random sampling

Inclusion And Exclusion Criteria

When a patient was admitted to the outpatient or emergency department with a medical or pediatric nonsurgical condition, their records were reviewed. On admission, a medical officer must have assessed the patient and noted the final diagnosis, which is also the discharge diagnosis. This was considered the main reason for admitting and treating the patients. There were no age restrictions for patient records. Records were excluded if the patient was admitted directly to the ward (rather than via OPD or ED) and had a surgical or obstetric condition. In addition, records were excluded when there was no evidence that a patient had been examined by a medical officer working with other team members, when the patient had left the ward, or when the patient had been referred to a regional referral hospital. Records of referred patients were excluded because it was assumed that a possible new diagnosis would be made at the referral centers. In contrast, an autopsy could be performed on deceased patients, which could lead to conflicting results. Incomplete records were excluded (if they were missing more than 30% of the variables of interest), if the medical officer failed to record the final discharge diagnosis, or if the patient was discharged against medical advice.

Data Collection

Data were collected by four trained research assistants who worked as medical records at two of the hospitals and were supervised by the first author. The data collectors received a two-day training course that included a pre-test of FY2018/2019 patient record tools at one of the study hospitals. An explicitly designed Excel-based form was used to collect the data.

Outcome Measures: The primary outcome of the quantitative study was patient misdiagnosis. The secondary outcomes are the factors (variables) related to misdiagnosis discovered by previous research (18–25).

Patient Misdiagnosis: If the initial diagnosis (admission) and final diagnosis (discharge) did not match, the patient's diagnostic status was classified as misdiagnosed. The diagnosis was considered correct if there was agreement or consistency between the admission diagnosis (initial diagnosis) and the discharge diagnosis (final diagnosis) (no misdiagnosis). In this study, too, a patient was considered to be misdiagnosed if the admitting diagnostician prescribed treatment without a definitive or preliminary (working) diagnosis. and based on previous studies of this type of misdiagnosis, the methods described herein were used. Previous research on diagnostic errors based on the difference between admission and discharge influenced the choice of this definition (26–30). Hautz et al. (2016) examined diagnostic differences between emergency departments and admissions departments and hypothesized that the differences were an indicator of diagnostic error (28). Kijima et al. (2018) used this method to examine the diagnostic accuracy of acutely ill elderly patients who presented to a primary care center and then admitted to an emergency department within four days (26). Avelino-Silva and Steinman (2020) reported on a study conducted in Brazil to identify diagnostic discrepancies between admission and discharge diagnoses, not all but the majority of which are due to errors (30). Similarly, a Japanese study examined the misdiagnosis of infection sites by comparing discrepancies between the initial site of infection and the final diagnosis (27).

The research assistant recorded both the admission and discharge diagnoses as they were written on the records during the record review. Later, two of the authors (SPK and RAO), both senior clinicians, assigned the ICD-11 codes using alphanumeric International Classification of Diseases Eleventh Version (ICD-11) software (31). For each diagnosis, the major diagnostic grouping, parent grouping, and individual disease condition digit numbers were recorded. When there was a difference of opinion about the codes, the two authors discussed and consulted extensively until they agreed on the code to be assigned.

The Secondary Outcomes:

Patient, disease or case-related, contextual, and health-system-related variables were considered to be associated with misdiagnosis (18–25). Age, sex, and whether the patient was new or old were among the extracted patient data. A new patient was defined as someone who made use of outpatient or inpatient healthcare services from the healthcare facility for the first time within the month prior to the admission currently being studied, and an old patient was someone who had previously visited the same hospital as an outpatient or inpatient within one month of the current admission. Disease-related variables extracted included the number of signs and symptoms noted at OPD/ED, the total number of final diagnoses made by the reviewing physician, and the type of disease the patient was finally diagnosed with. The diseases were classified as either common or uncommon admission diseases. Common admission diseases are the 20% of the conditions or diseases accounting for 80% of admissions. To do this, all final diagnoses were ranked according to the frequency of hospital admissions and the diseases responsible for 80% of admissions were identified using the Pareto analysis (32). These were the diseases classified as common admission diseases, while the others were classified as uncommon admission diseases. Malaria, Pneumonia, Severe Anaemia, Peptic Ulcer Disease, Gastroenteritis, Bacteraemia, Urinary Tract Infection (UTI), Septicaemia, Hypertension, Enteric Fever, Sickle Cell Disease with Crisis, Diabetes Mellitus, and Acute Malnutrition were among the common illnesses. We also determined and gathered information on whether the initial diagnosis was made as a result of any imaging or laboratory investigations. We also determined whether the patient had an underlying disease or condition, as reported by the ED/OPD diagnostician.

We also extracted variables related to admission time, such as patient admission time and day of the week. The day was divided into three periods: daytime (8:00 am to 2:59 pm), evening time (3:00 pm to 10:49 pm), and night time (8:00 pm to 2:59 pm). (11:00 p.m. to 7:59 a.m.). The day of admission was either a weekday (Monday to Friday) or a weekend (Saturday and Sunday). One of the health system related factors was the referral status of the patients with two classifications of referred or non- referred patients. A referred patient was someone who had received treatment from a lower-level private or public healthcare facility but was later transferred to the hospital for further treatment, while a non-referral patient had not received treatment from a healthcare facility and had not been transferred to the hospital.

Quantitative Data analysis: The data were analyzed with StataIC version 15.0 by DEK and SPK. Frequencies, proportions and confidence intervals (CIs) were reported for descriptive characteristics. The chi-square test was used in bivariate analysis to determine factors associated with misdiagnosis. Multivariable regression analysis was used to fit a model that determines the factors associated with patient misdiagnosis while controlling for confounders. The level of significance was set at p 0.05.

Qualitative Study

The qualitative study was conducted following the completion of the first phase, which included record review and data analysis from September 2020 to April 2021. Phase 2 (qualitative data collection and analysis) took place between April and June 2021. Using a phenomenological approach, interviews in the qualitative phase with health professionals involved in diagnostic processes in hospitals. Every healthcare worker is expected to be involved in a misdiagnosis incident at some point in their career (33). As a result, the decision to collect data from a small group of respondents who are involved in the diagnostic process (phenomenon) under investigation was made due to its ability to collect rich and detailed data from respondents who have firsthand experience with the diagnostic process that results in a different outcome than expected (34). This was supported further by the idea that in the healthcare setting, learning from the experiences of others greatly aids in improving diagnostic performance.

Purposive criterion sampling was used to select respondents for in-depth interviews based on pre-established criteria related to the study's research objectives (35). Participants were considered for in-depth interviews if they were working at one of the study hospitals at the time of the study and had at least two years of clinical officer or medical officer experience in their respective jobs. After obtaining written informed consent, twelve qualitative interviews were conducted with 12 participants, six clinic staff and six physicians. Participants included Clinical Officers and Medical Officers. These two cadres are tasked with performing diagnostic roles in Uganda's health care system (36–38). A diagnostician is a healthcare professional who is responsible for carefully taking the medical history, conducting examinations, ordering examinations, interpreting the examinations, and using all information from these procedures to identify the patient's disease (39).

Clinical officers make the initial (admission) diagnosis in the outpatient or emergency department, and medical officers review admitted patients on the ward to make the final diagnosis (36). Because of their superior training compared to the clinical staff, medical staff are believed to have better diagnostic expertise than clinicians and hence they provided critical perspectives regarding the interface with the patients they screened (38). In addition, when reviewing admitted patients, they collaborate and consult with other team members such as nurses, radiologists, laboratory technicians and colleagues to arrive at a definitive diagnosis, which clinicians may not be able to do at the first interface with the patient, if he or she has just arrived in the OPD or ED. As a result, this process leads to a more reliable and definitive diagnosis than clinicians have likely made.

In order to recruit the interviewees for the qualitative interview, the medical officers were contacted via the medical superintendent and the clinic officers via their supervisor. All study participants were informed that they will participate in a doctoral study, the results of which will be published in a peer-reviewed journal. The total number of respondents was determined by the saturation point (the point at which no additional information could be gleaned from additional interviews). The tool was pretested among some diagnosticians in Entebbe hospital, which was not one of the study hospitals. PA and AAA, both senior health systems researchers, evaluated the interview guide for face validity, as did RAO, a Senior Clinical Officer who is also a senior quality of care and patient safety and health systems researcher. SPK collected data until saturation using a similar pretested in-depth interview guide that was designed based on quantitative results and study objectives. The guide began with broad questions that sought perceptions Participants' perceptions and beliefs about the occurrence of misdiagnosis in hospitals, as well as the factors commonly thought to contribute to patient misdiagnosis, were investigated using broad health system building blocks. Later questions focused on the relationship between misdiagnosis and (a) the time of admission, (b) the number of morbidities, (c) referral status, patients' demographics such as (d) age and (e) sex, and disease-related factors such as (f) the presence of underlying diseases, and (g) uncommon diseases.

The interviewer used the bracketing technique to avoid expressing his views on the patient's misdiagnosis while remaining flexible and allowing the conversation to flow and connect the interview to the quantitative results (40, 41). Exploratory questions were used to get more information about the answers given. Notes were taken during the interviews in order to be able to form a brief opinion on each interview. The interviews were tape-recorded with the consent of the participants. Twelve participants (six clinicians and six medics) from Kiboga, Nakaseke and Mityana hospitals, one from Gombe and two from Kayunga hospitals reached saturation. Although it was deemed necessary to collect more data from participants at Gombe and Kayunga hospitals to better understand their perceptions, this was not possible due to the COVID-19 pandemic restrictions prohibiting in-person interviews, which the identified participants did prefers.

Qualitative Data Analysis

SPK listened to all the recordings, familiarized himself with the information, and supported the transcriptions with notes made during the interviews. After thorough reading of the transcripts, SPK and RAO used deductive thematic analysis to analyze the data. The data analysts used the Excel sheet to collect the diagnosticians' experiences, explanations, meaning and reality, which they then tailored to the four previously identified themes (42). The main issues influencing patient misdiagnosis were (1) disease-related, (2) patient-related, (3) context-related, and (4) health-system-related factors. For each of the main themes, sub-themes were identified from which the codes were developed. Extracts found to match specific codes were obtained and used to support the idea presented and to express the main message in the data, as guided by previous research (43–46). PA and AAA then independently reviewed and refined the analyzed data for quality, robustness, and credibility

Integration Of Mixed Methods

Following the multivariable analysis of the quantitative data, the findings of the quantitative and qualitative studies were combined during the interpretation stage of this study to clarify the quantitative study's findings.

Characteristics of patient records reviewed

Of the 12,345 records retrieved, 5,832 met the inclusion criteria. By simple random sampling, 2431 records were obtained, which were then analyzed. Records of children aged 0 to 9 years accounted for up to 41.3% (n = 1003) of all records (n = 1003). Female patients outperformed male patients by a narrow margin of 51.9–48.1%. The vast majority of patients (96.3%, n = 2341) were admitted as new cases (first treated within the previous month). The majority of patient diagnoses (84%, n = 2041) were made after some pre-admission investigations. Up to 13% of patients (n = 317) presented with more than three complaints. At the time of initial diagnosis, the vast majority of patients (84.5%, n = 2055) reported no known underlying disease. The majority of patients (81.9%, n = 1991) had a single disease diagnosis recorded as the definitive (discharge) diagnosis, while 1861 (76.6%) had a definitive (discharge) diagnosis(s) recorded as common admission diseases were classified. 2076 (85.4%) of the patients were admitted during the day and 81.4% (n = 1976) on weekdays (Monday to Friday). Up to 5.4% of patients (n = 131) were referred from subordinate health care institutions (Table 1).

Table 1

Characteristics of study variables
Variable		Frequency (f)	Percentage (%)
Hospital (N = 2431)	Kiboga	481	19.8
	Nakaseke	489	20.1
	Mityana	485	20.0
	Gombe	489	20.1
	Kayunga	487	20.0
Age category (N = 2431)	0–9	1003	41.3
	10 to 19	408	16.8
	20–29	315	13.0
	30–39	199	8.2
	40–49	141	5.8
	Above 50	365	15
Gender	Female	1262	51.9
Gender	Male	1169	48.1
Referral status of the patient	Referred	131	5.4
Referral status of the patient	Not referred	2,300	94.6
Type of patient	Old patient	90	3.7
Type of patient	New patient	2341	96.3
Had initial diagnosis been made after investigations	No	390	16.0
Had initial diagnosis been made after investigations	Yes	2041	84.0
Number of presenting complaints	More than 3	317	13.0
Number of presenting complaints	1–3	2,114	87.0
Did patient have an underlying health condition	Yes	376	15.5
Did patient have an underlying health condition	No	2055	84.5
The number of morbidities	One disease	1991	81.9
	Comorbidity (2 final diagnoses)	390	16.0
	Multimorbidity (3 or more final diagnoses)	50	2.1
Time of admission	Day time	2,076	85.4
	Evening	223	9.2
	Night	132	5.4
Period of the week patient was admitted	Weekday	1976	81.3
Period of the week patient was admitted	Weekend	455	18.7
How common is the diagnosis made	Combination of common and uncommon disease	144	5.9
	Only common disease	1861	76.6
	Only uncommon disease	426	17.5
Diagnostic status	Not misdiagnosed	2208	90.8
Diagnostic status	Misdiagnosed	223	9.2

Participants For The Qualitative Analysis

Twelve respondents were interviewed face-to-face, six of whom were clinical workers and the other six were medical workers. The participants in the study have gained a variety of work experiences. There were only three female participants. Table 2 summarizes the characteristics of the qualitative study participants.

Table 2

Study participants and their demographics
Rank category of profession	Age	Sex	Years of work experience
Clinical Officer 1	54	Male	30 + years
Clinical Officer 2	41	Male	15 years
Clinical Officer 3	48	Female	20 + years
Clinical Officer 4	46	Male	15 years
Clinical Officer 5	44	Female	20+
Clinical Officer 6	57	Male	30+
Medical Officer 1	41	Male	15 years
Medical Officer 2	39	Male	10 years
Medical Officer 3	55	Male	25 years
Medical Officer 4	49	Female	15 + years
Medical Officer 5	48	Male	15 + years
Medical officer 6	45	Male	20 + years

Primary Outcomes

In total, 223 people (9.2%; 95% CI: 8.1–10.3%) were misdiagnosed. Nakaseke hospital accounted for 27.8% (62) of the 223 misdiagnosed cases, while Kiboga hospital had the lowest (14.4%) (Table 3).

Factors Associated With Misdiagnosis At Bivariate Analysis

At bivariate analysis, misdiagnosis was associated with hospital (p = 0.005), age (p0.001), gender (p < 0.001), patient referral status (p < 0.001), patient type (p < 0.001), and whether the initial diagnosis was made after some investigations (p = 0.001). Other factors included the presence of an underlying disease (p0.001), the number of morbidities (p0.001), the time of admission (p < 0.001), the week of admission (p = 0.026), and the frequency with which the diagnosis was made (p < 0.001).

Table 3

Cross-tabulation of independent variables and misdiagnosis of in-patients based on the Chi-square test statistic with corresponding p-value in general hospitals of Central Uganda
Variable		Total	Diagnostic status				Chi2	p-value
Variable		n	Not Misdiagnosed (f)	%	Misdiagnosed (f)	%	Chi2	p-value
Hospital (N = 2431)	Kiboga	481	449	20.3	32	14.4	14.7551	0.005
	Nakaseke	489	427	19.3	62	27.8
	Mityana	485	434	19.7	51	22.9
	Gombe	489	445	20.2	44	19.7
	Kayunga	487	453	20.5	34	15.3
Age category (N = 2431)	0–9	1003	990	44.8	13	5.8	196.7705	0.000
	10 to 19	408	382	17.3	26	11.7
	20–29	315	274	12.4	41	18.4
	30–39	199	165	7.5	34	15.2
	40–49	141	117	5.3	24	10.8
	Above 50	365	280	12.7	85	38.1
Gender	Female	1262	1,186	53.7	76	34.1	31.2745	0.000
Gender	Male	1169	1,022	46.3	147	65.9	31.2745	0.000
Referral status of the patient	Referred	131	101	4.6	30	13.4	31.3172	0.000
Referral status of the patient	Not referred	2,300	2,107	95.4	193	86.6	31.3172	0.000
Type of the patient	Old patient	90	66	3.0	24	10.8	34.3277	0.000
Type of the patient	New patient	2341	2,142	97.0	199	89.2	34.3277	0.000
Number of presenting complaints	More than 3	317	284	12.9	33	14.8	0.6694	0.413
Number of presenting complaints	1–3	2,114	1,924	87.1	190	85.2	0.6694	0.413
Had initial diagnosis been made after investigations	No	390	337	15.3	53	23.8	10.8536	0.001
Had initial diagnosis been made after investigations	Yes	2041	1,870	84.7	170	76.2	10.8536	0.001
Whether patient had an underlying disease	Yes	376	296	13.4	80	35.9	78.2065	0.000
Whether patient had an underlying disease	No	2055	1,912	86.6	143	64.1	78.2065	0.000
The number of morbidities	Multimorbidity (3 or more diagnoses)	50	33	1.5	17	7.62	44.2352	0.000
	Comorbidity (2 diagnoses)	390	343	15.5	47	21.08
	One disease	1991	1,832	83.0	159	71.3
Time of admission	Day time	2,076	1,908	86.4	168	75.3	42.7234	0.000
	Evening	223	201	9.10	22	9.9
	Night	132	99	4.5	33	14.8
Period of the week patient was admitted	Weekday	1976	1,806	81.9	169	75.8	4.9310	0.026
Period of the week patient was admitted	Weekend	455	400	18.1	54	24.2	4.9310	0.026
How common is the diagnosis made	Had both common and uncommon disease	144	116	5.3	28	12.6	228.2944	0.000
	Only common disease	1861	1,780	80.6	81	36.3
	Only uncommon disease	426	312	14.1	114	51.1

Factors associated with misdiagnosis in the multivariable analysis included; the patient's age, gender, referral status, number of morbidities, time of day admitted, and the type of disease (es) the patient had as the final diagnosis (Table 4).

Table 4

Multivariate logistic regression on factors associated with inpatients misdiagnosis in general hospitals of Central Uganda (N = 2431)
	Variable	Misdiagnosed n %		Odds Ratio	P value	95% CI
Hospital	Kiboga	32	14.4	Reference
	Nakaseke	62	27.8	1.95	0.01	1.17 3.25*
	Mityana	51	22.9	1.40	0.213	0.83 2.37
	Gombe	44	19.7	0.98	0.956	0.57 1.69
	Kayunga	34	15.3	0.90	0.722	0.51 1.59
Age	0–9	13	5.8	Reference
	10 to 19	26	11.7	4.61	0.0000	2.30 9.25*
	20–29	41	18.4	8.15	0.0000	4.18 15.89*
	30–39	34	15.2	8.12	0.0000	3.99 16.54*
	40–49	24	10.8	7.88	0.0000	3.71 16.73*
	50+	85	38.1	12.14	0.0000	6.41 23.01*
Sex	Female	76	34.1	Reference
Sex	Male	147	65.9	1.89	0.0000	1.35 2.64*
Referral status	Referred	30	13.4	Reference
Referral status	Not referred	193	86.6	0.51	0.011	0.31 0.86*
Case type	Old patient	24	10.8	Reference
Case type	New patient	199	89.2	0.94	0.85	0.51 1.75
Lab tests done	No	53	23.8	Reference
Lab tests done	Yes	170	76.2	1.18	0.41	0.79 1.76
Underlying disease	Yes	80	35.9	Reference
Underlying disease	No	143	64.1	0.63	0.015	0.43 0.91*
Number of morbidities	One disease	17	7.62	Reference
	Comorbidity (2 diseases)	47	21.08	1.49	0.1	0.93 2.38
	Multimorbidity	159	71.3	4.71	0.001	1.91 11.65*
Time of admission	Day time	168	75.3	Reference
	Evening	22	9.9	1.26	0.10	0.74 2.14
	Night	33	14.8	3.00	0.000	1.81 5.02*
Period of the week	Weekday	169	75.8	Reference
Period of the week	Weekend	54	24.2	1.23	0.279	0.84 1.79
How common the diagnosis is	Combination of common and uncommon disease	28	12.6	Reference
	Only common disease	81	36.3	0.55	0.083	0.28 1.08
	Only uncommon diseases	114	51.1	2.57	0.008	1.28 5.18*

Diseases/conditions Involved In Misdiagnosis

Table 5 shows the ICD-11 codes of the initial (misdiagnosed) and final (correct) conditions or diseases.

Table 5

The ICD-11 coding of the initial and final diagnoses for the misdiagnosed patients (n = 223)
Initial (wrong) diagnosis	Final (right diagnosis)	Initial (wrong) diagnosis	Final (right diagnosis)
1. QA1C	MG24.1	111. DA42.Z; MB24.3	1A40.Z
2. MA15.0; FA20.Z	8B20	112. MG40.Z	MG45.Z
3. 1F40.Y, 1A07.Z, MA15.0, DA61	GB51	113. CA40.Z; 1B70.Z; FA05	CA22.Z; BD10&XT5R; 8A61.3Y
4. 1F40.Y; DA42.Z	1A07.Z; DA61; 1A40.Z	114. FA8Z; CA45; MF50.2Y	FA80.9; CA20.Z;5B81.01
5. QA1C	1B95	115. QA1C	MB21.4; 3A9Z
6. 1A90; AA8Z XT5R	1E91.Z; 1C62.3; 1F2Z	116. CA20.Z; GCO8.Z	1B10.1; 3A9Z; 5B54
7. IC62.Z	GA90	117. BC4Z; DA61	BD10
8. 1A40.Z	1A1Z	118. GB01.1	GC00.Z; GB02.Z
9. CA23	CA22.0	119. QA1C	8A80.Z
10. QA1C	1F40.Z	120. GB41; BD10	GB40; 9C80.0
11. QA1C	5A41; 6A70.Z	121. 1B98; DA61; DA42.Z, 4A84.Z	1F57.Z
12. GA05.Z, GC00.Z	GCO8.Z	122. QA1C	5B7Z; 1B10.0; 5C70.0
13. CA40.07	1B10.0	123. QA1C	CA40.Z&XT6S&XY69
14. PA78&XE11V	1B70.Z	124. MD81.3; ME24.90; 1B12.7	1A40.Z; DC50.0
15. QA1C	DC12.Z&XT8W	125. QA1C	1E50.1; DA61
16. QA1C	1A40.Z	126. QA1C	8B20; BA00.Z
17. BD11.0	BD13	127. GCO8.Z; GB51	GC00.Z
18. QA1C	1F40.Y	128. 1A40.0	ME05.1
19. QA1C	1F40.Y	129. AB0Z	AA91.Z
20. QA1C	5A41	130. QA1C	MA15.0
21. QA1C	1F40.Y	131. 3A9Z	8B20; 8A6Z
22. 3A9Z	DA61	132. BD1Z	5DOY; BA00.Z
23. DA61; 1F40.Y	1A07.Z; 1B95; CA45	133. 6C40.3; 5C70.0	5B5A.10; 5C70.0; CA40.Z
24. QA1C	DA61	134. GCO8.Z	1F40.Y
25. QA1C	MA15.0	135. QA1C	1F45; GC00.Z
26. 1F45; GCO8.Z	1B10.1	136. QA1C	DA61
27. QA1C	CA40.Z&XT6S&XY69	137. 1F45	GCO8.Z
28. 1F40.Y; 1A07.Z; 1E50.Z	GB51	138. 1F45	1A40.Z
29. 6B01; 6C20.Z	6B60.Z	139. 1F40.Y	1A40.Z; BA00.Z
30. NE60	EB13.0	140. QA1C	1F45
31. GB60.Z; 4B40; 2C94.Z	GB61.Z	141. CA22.Z	DA61
32. ME84.2Z; DA61; NB9Z	FA8Z	142. QA1C	1A07.Z; 3A9Z
33. BC20.1; BB01.Z	BD11.Z	143. 1B10.Z	CA40.Z
34. 1C62.Z	1C62.3	144. 1A1Z	DA61
35. 1F45; 1C41; 1A36.00; 1A07.Z	BA00.Z	145. DA42.7	1F45; DA61
36. 1A07.Z	DC50.Z	146. 6C40.3; 3A9Z	NA0Z; 6C40.3
37. 1B10.1	CA40.Z&XT6S&XY69	147. QA1C	1F40.Y
38. 1A40.0	1A07.Z	148. QA1C	1F40.Y
39. QA1C	IF40.Z	149. CA45	1B10.1
40. GA05.Z	DA61	150. 3A9Z	5A21.Z; BA00.Z
41. MA15.0	DA61	151. 8C01.3; FA92.Z	FA2Z; GCO8.Z
42. 1F45; 5A41	6A20.0Z; 8A6Z	152. CA40.Z	GC00.Z; 3A9Z; ME05.0
43. QA1C	BA00.Z	153. BA00.Z; 8C0Z	BA01
44. DA61	1B95	154. 6A2Z; 5A14	BA03; 5A14
45. 1A40.Z; DA61	DA42.Z	155. BD10; CA20.Z	BA6Z
46. QA1C	DA61	156. GB6Z; DA61; 5B5A	1A40.Z; 1A07.Z; DB94.Z; MA18.0
47. 6C4Z	6C4G.70; BA00.Z	157. DA42.Z	DA61; 1A07.Z
48. QA1C	6D10.Z	158. 1F45; MA15.0	1B10.1; 1F23.0
49. QA1C	5A14	159. CA45	GB6Z; GB61.Z; DA61
50. 8A68.Z	8A6Z	160. 6C2Z	1A40.Z; 6A2Z
51. QA1C	1F45	161. QA1C	1A40.Z; CA09.Z
52. QA1C	GCO8.Z; FA92.Z	162. QA1C	ME05.1
53. 5A14; BA00.Z; BC4Z	5A21.0	163. QA1C	CA40.Z&XT65
54. 4A84.Z	CA71.0	164. 1B40.Z	1A07.Z
55. 2C13.0	DD71.Z	165. DA61	6A60.9
56. 1F40; CA71.0; 5A22.Z	CA40.Z&XT65	166. 1F40.Y; MA15.0	FA8Z
57. DA61; BA00.Z; 3A9Z	GB5Y	167. 1F40.Y	MA15.Y
58. BA00.Z; 1F45	BD90.Y	168. 8B20; MA15.0	5A41
59. CA40.Z	8C03.0; BA00.Z	169. DA61	1C41
60. 1F40	1A07.Z	170. 1B10.1; BD10	BA01; CA07.0
61. CA40.Z	CA22.Z; 1B70.Z	171. 5C70.0; 1B10.1	CA45
62. QA1C	1A07.Z	172. 5C70.0; 1C41	1A40.Z
63. CA45; 5C70.0	1A07.Z; CA40.Z; 1A40.Z	173. CA40.Z; CB03.4	MD11.6
64. BD10; GC2Z; DB98.7Z	BD13; BB01.5	174. MC81.2	6D71; BD11.Z
65. 1D01.10	GCO8.Z; 8B42; 1A07.Z	175. DA61	DA42.Z; 1E90.0
66. 5B7Z; 5C70.0	5B52&XS25; 1A40.Z	176. CA40.Z; 1C60.Z	1C62.3; 3A9Z
67. QA1C	1A40.Z	177. BD10	MA15.Y
68. 5A11	DA42.Z; GCO8.Z; 1F45	178. CA07.0; 1F23.0	CA40.Z; DA61
69. MB24.5	1A62.0; 2B31.20	179. CA07.0	BA00.Z; CA40.Z; GCO8.Z; 1A07.Z
70. CA23.30	CB03.Z	180. ME05.1	1A40.Z; DA61
71. MG45.Z;1F45; GCO8.Z	DA22.Z; DA61	181. 6D85.3; 1B10.1	1D01.10; CA07.0
72. BA00.Z	BE2Y	182. DA61; 1A07.Z; DB94.3	DB94.3; 3A9Z; GCO8.Z
73. 6C40.Z; 5A14	1C8E; 5A21.Z	183. CA40.Z; 1B10.1	CA07.0; 6A60.9
74. CA20.1Z; 8B20	CA22.Z	184. 5A11; DA42.Z	GCO8.Z; DA61
75. MA15.Y; 1B10.1	DA42.82	185. 5A11	DA61
76. MA15.Y	BD10	186. CA07.0; 5C70.0	CA40.Z
77. QA1C	GCO8.Z	187. 1A07.Z	MA15.0
78. 1C41	8A80.Z; MA15.Y	188. DA61	DC50.Z
79. 3A9Z	MD31; DA61	189. QA1C	1F40.0
80. DA42.Z; CA07.0; 1B10.1	PD03	190. QA1C	1B10.1
81. QA1C	QA14	191. 1A40.Z; MA15.0	ME24.9Z
82. 4B23; 1B10.1; GCO8.Z	CA40.Z	192. 6A20.Z	9C83.13
83. CA40.Z	BD10	193. MA15.0	MA15.Y
84. CA45; 5C70.0	MA15.Y	194. BA00.Z	BD10
85. 5C70.0; 5B52&XS25	DA22.Z	195. GC00.Z	ME10.02
86. MD81.3	DA61	196. QA1C	2C77.Z
87. QA1C	MA15.Y	197. CA23.30	1B10.1
88. 1C41; DA42.Z	CA45	198. BA00.Z; DA61; CA40.Z; CB27	BA01; 5A11; CA22.Z
89. MA15.0	MA15.Y	199. MA15.0	5A11
90. QA1C	MG22	200. DA61	BA03
91. 1C62.Z; 1F23.2; CA07.0	CA45; 1D01.10, 1B10.1	201. MG40.Z	1F40.0
92. DA61; DB93.1	DB95.Z	202. MF50.3; 1A6Z	BE2Y
93. CA07.0	MA15.0	203. BE2Y; 1F40; CA40.Z	1F57.Z
94. 3A51.2	1F40.0; GCO8.Z	204. MA15.0	1D01.10
95. KA60	KB8Z	205. 5A11	MA15.0
96. MA15.Y	KA87.Y	206. 5A11	DA42.Z
97. MA15.Y	KB06	207. 8A81.Z	6A7Z
98. DA61; BA2ZZ; 3A9Z&XS25	2B33.1	208. CA07.0	5B7Z
99. GB40	GB61.Z	209. GB40	BD10
100. QA1C	CA40.Z	210. 1C62.Z	6D85.3
101. DA61; 1C41; CA07.0; GCO8.Z	8B42	211. 5C70.0	8B20
102. MC81.2	BD10	212. QA1C	1F45
103. QA1C	CA45	213. 1F40.Y	5C70.0; 3A9Z
104. ME05.1; 5B7Z	GB61.Z; 1A40.Z	214. 1A07.Z	1E50.Z
105. ME04.Z	BD10	215. GB02.Z; GCO8.Z	1A6Z
106. 1F40.Y	GB51	216. CA00	CA08.1Z
107. 5A11	5B57.Z	217. QA1C	MA15.0
108. MA15.0	1D01.10	218. GB40	GB61.Z
109. QA1C	1A40.Z	219. QA1C	CA40.Z
110. 1A07.Z	GA34.3	220. DA61; 1C41; CA07.0; GCO8.Z	8B42
111. 3A9Z&XS25	3B4Z	221. MC81.2	BD10
112. MA15.0	CA22.Z	222. MC81.2	6D71; BD11.Z
113. QA1C	1F23.0	223. 4A84.Z	CA71.0

The contextual factors associated with misdiagnosis

Patients’ Misdiagnosis And Night Admission

Table 4 shows that night duty was significantly associated with patient misdiagnosis (aOR 3.0, 95% CI: 1.81–5.02, p < 0.01) compared to patients admitted to other shifts. Participants attributed this finding to several reasons, as discussed below.

Clinicians Fatigued At Night

The high likelihood of patients being misdiagnosed at night is fatigue which impairs clinical cognition:

‘There is only one clinician on duty at night, who may be tired and will likely only probe a little about the symptoms and signs, rather than examine and delve deeper into the patient's condition. (Clinical Officer 5).

On the other hand, it has been pointed out that some clinicians have chosen to work chronic night shifts, which predisposes them to fatigue and cognitive difficulties and makes them prone to error.

Difficult To Do Investigations And Examination At Night

The difficulties in conducting patient investigation at night stem from problems such as the inability to perform laboratory or radiological studies due to power outages and the difficulty in reaching laboratory personnel who are normally only on call at night for emergencies especially for patients who may need blood transfusion:

‘…We do not have access to laboratory tests at night, so we conduct investigations in the morning to arrive at a final diagnosis. the laboratory only handles emergencies and blood transfusions at night (Clinical Officer 4).

It has also been found that access to patient screening equipment such as blood pressure monitors, stethoscopes, thermometers, weighing scales and clinical guides is difficult at night in some hospitals as they are kept under lock and key for fear of theft.

'To avoid theft, some instruments, such as the sphygmomanometer, pulse oximeter, and thermometer, are kept locked at night. As a result, you might not have access to them when you need them to examine the patient. Thus, the clinician admits the patient based solely on symptoms and signs, increasing the chances of misdiagnosis (Clinical Officer

Patient Management At Night Is Centred On Stabilizing The Patient

Physicians on night duty tend to focus more on stabilizing the patient's condition, regardless of the accuracy of the patient's diagnosis. As a result, it was common to find either no diagnosis or a tentative diagnosis assigned to the patient based on less history, examination, and investigations. For example, one interviewee explained:

‘… We do not provide a final diagnosis at night. We only manage to stabilize that patient; then, in the morning, we conduct investigations and arrive at a final diagnosis. If you select those at night, you will discover that some of them are missing a provisional diagnosis, depending on the type of Clinical Officer who was present. (Medical Officer 2)

Difficult To Consult Colleagues At Night When Challenged

Participants reported that clinicians who work nights find it more difficult to consult with others about a difficult condition than they do during the day. Patients presenting to the outpatient clinic for non-surgical emergencies are less likely to receive appropriate diagnosis because the medical officers whom clinicians typically consult when faced with a challenging condition primarily prefer to be consulted about surgical, accident and pregnancy related emergencies at night.

‘There is a high likelihood that those working during the night will not provide an accurate diagnosis. There is only one doctor on duty at night, and his or her priority is likely to be maternal emergencies, accidents, or burns. The majority of the cases you mentioned as the most frequently misdiagnosed are for internal medicine, which is unlikely to be supported by Medical Officers. So, if a clinical officer is unsure and wishes to consult with the Medical Officers, he or she must wait until dawn (Medical Officer 1)

Patient related factors and misdiagnosis

Male Patients Are More Likely To Be Misdiagnosed

According to the results in Table 4, male patients were 89% more likely than female patients to be associated with misdiagnosis (aOR = 1.89, 95% CI = 1.35–2.64, p < 0.01). This finding can be explained by the fact that the healthcare system in Uganda has historically focused mainly on maternal and child healthcare, with little attention given to male health. Consequently, men tend to display inadequate health-seeking behaviors, which are manifested by delayed hospital reporting when their conditions have already worsened and become challenging to treat, as observed by one clinician;

Men are not exposed to the hospital system. They cannot give themselves fully to the clinician. On the other hand, the females are always interfacing with the clinicians in the facility so they are open with them’ (Clinical Officer 2).

The participants also noted that male patients had less interaction with the healthcare system compared to female patients, resulting in their unfamiliarity with how to actively engage in clinical decision-making, particularly in responding to questions posed by diagnosticians during consultations:

‘‘Whatever they want, they come with their own perceived way of how they will say it. “Now, they should do this to me if not, I will walk away”. The female will give more relevant history than the male when they are asked during examination’ (Clinical Officer 3)

Furthermore, some participants stated that, in contrast to the majority of female patients, certain male patients are uninterested in consultations and the diagnostic process. As a result, they will provide ambiguous responses that will not directly guide the diagnostic process, leading to misdiagnosis.

‘Female patients will openly express their complaints, including their secrets. You must exercise extreme caution when dealing with men. They are more subdued than female patients’ (Medical Officer 5)

Patients’ Age

Patients aged 10–19, 20–29, 30–39, and 40–49 were more likely to be misdiagnosed 2.3 times (aOR = 2.3, 95% CI: 2.3–9.25, p < 0.01), 8.2 times (aOR = 8.2 95% CI: 4.18–15.89, p < 0.01), 8.12 times (aOR = 8.12 95% CI 3.99–16.54). Several reasons were given during the qualitative interviews as to why diagnosing diseases in children is often easier than in adults. Some of these reasons include the fact that, unlike adults, children rarely make false claims about their illnesses, there are fewer pathologies that are easily identified in children, and the complexity of diseases in adults makes diagnosis more difficult.

Children Less Likely To Falsify Illnesses

According to the participants, children, unlike adult and elderly patients, do not usually lie about their illnesses, making them less prone to misdiagnosis:

‘Adults may conceal their symptoms, but children do not. With an adult, you usually have to sift through a lot of information to get to the truth’. You will know if a child has a fever, if he is vomiting, if he is playing, and if he is not playing (Medical Officer 5).

Fewer Pathologies In Children Means Fewer Incorrect Diagnoses

Respondents indicated that children are more likely to receive a correct diagnosis compared to older patients because when diagnosing children, diagnosticians need to consider a smaller range of diseases compared to adults:

‘Morbidity in children under the age of five, I believe, is simply a collection of a few easily predictable diseases. For example, if a child has a fever, the clinician is almost certainly thinking it is malaria, which is the most common here, but it could be bacteraemia, or if a child has diarrhoea, they are well aware of it. Pneumonia is diagnosed when a child has a cough, fever, and difficulty breathing. Then there are anaemia and sickle cell disease issues in high-risk areas’ (Medical Officer 4)

Complexities Of Adult Conditions

Participants also noted that making accurate diagnosis is easier with children since their conditions tend to be less complex compared to those of the adults:

‘Unlike adults, who may present with complex ailments that require some level of expertise to correctly diagnose, a paediatric patient is more likely to present with either of dehydration, pneumonia, malaria, or a sickle cell disease crisis, all of which are easy to diagnose’ (Medical Officer 3).

Due to the complexity of the diseases in adults, a thorough anamnesis and examination is usually not possible due to the working conditions and the limited time in outpatient clinics or emergency rooms.

Disease Related Factors And Misdiagnosis (Dup: Abstract ?)

Presence of underlying diseases and misdiagnosis

Table 4 also shows that the absence of underlying diseases is associated with a 37% lower risk of misdiagnosis (aOR = 0.63; 95% CI: 0.43–0.91, p = 0.015). Participants in the qualitative study explained that mentioning underlying conditions could confuse clinicians, causing them to focus too much on the underlying condition and overlook other potential causes of the problem, eventually leading to a patient misdiagnosis. Furthermore, diagnosticians may be perplexed by patients who have extensive knowledge of underlying conditions and may overwhelm clinicians with their knowledge.

You cannot have come to the hospital because of all three morbidities. You must have been aroused to come to the hospital to seek treatment because of a specific morbidity. But the problem is when a patient comes here, he/she will tell the clinician that " You know I also have diabetes, I have pressure, I have this". As a result, even clinicians may be perplexed, and they may fail to focus on what brought this patient to them (Medical Officer 1).

Other participants stated that misdiagnosis can be significantly associated with attendants who are well-informed about an unconscious patient with comorbidity and sway the diagnostic decision-making process in the direction of the comorbidity. Such attendants may mislead diagnosticians, resulting in a false diagnosis.

Sometimes It is possible especially when unconscious patients in critical condition are brought in by desperate well-wishers or relatives who tend to know and exaggerate only about the patient’s underlying diseases. It is such patients who are likely to be misdiagnosed if the non-critical clinician is inclined to follow the attendants’ line of thought (Clinical officer 4).

Mentioning an underlying disease may also draw the clinician's attention away from other potential conditions and towards the complications of the underlying disease.

Most clinicians prefer to think about complications from the same chronic illness rather than a new illness. So, it all comes down to the Clinician's ability to detect any other condition discovered during the process (Medical Officer 1).

Multimorbidity And Misdiagnosis

According to the findings in Table 4, patients with multimorbidity are approximately five times more likely to be associated with a misdiagnosis than patients with one morbidity (aOR = 4.71; 95% CI: 1.91-11 .65, p0.01). Participants in the qualitative study explained that in cases of multimorbidity, especially when patients present with multiple symptoms and signs, diagnosticians are more likely to focus on the dominant symptoms and signs and may overlook others that could have revealed the true illness. The following explanation is a good example of this;

That one is always there and somehow connected to the next syndrome. When there are many patients and a patient has multimorbidity, such as cough, flu, sores in the nostrils, or oral sores, the clinician is likely to simply say there is infection and prescribe an antibiotic to clear it. However, this person has oral candidiasis and tonsillitis, and you are treating them with amoxicillin, which will not clear the fungal infection that you have misdiagnosed [Clinical Officer 6].

Rare Or Uncommon Diseases

Based on Table 4, patients who received treatment for uncommon hospital admission illnesses had approximately three times higher chances of being linked with misdiagnosis compared to those who were treated for both common and uncommon hospital admission illnesses aOR = 2.57; 95% CI: 1.28–5.18, p < 0.01). The health workers acknowledge that diagnosing uncommon or unfamiliar diseases is a challenging task. They argue that when a clinician confronts a rare illness, they are more prone to ascribe a diagnosis that they are familiar with and whose symptoms resemble the rare disease. One participant mentioned that:

‘Clinicians have been taught in school with emphasis on the signs and symptoms that point to a specific diagnosis that is common in their area. When patients present with abdominal or any chest pain near the gastric region without cough, a clinician with limited knowledge of cardiac problems believes it is peptic ulcers or typhoid and may overlook other possibilities (Medical Officer 2).

Health system factors associated with misdiagnosis

After controlling for other variables, the results revealed that the odds of being misdiagnosed were two times higher in Nakaseke Hospital patients than in Kiboga Hospital patients (adjusted Odds (aOR) = 1.95, 95% CI: 1.17–3.25, p = 0.01). In addition to what has already been stated, health system factors have been used to explain why misdiagnosis may occur not only at Nakaseke hospital but also in other hospitals.

Human resources for health and misdiagnosis

Scarcity Of Human Resources

Participants highlighted the shortage of diagnosticians, which contributes to high work pressure in outpatient clinics and emergency departments. The desire to care for as many patients as possible likely overtook the desire to care for each patient effectively, resulting in physicians failing to adequately screen patients, leading to patient misdiagnosis:

‘There are times when a single clinician is responsible for too many patients. When he or she becomes tired, he or she is bound to fall victim to the "next in line syndrome" and fail to properly examine patients” (Clinical Officer 4).

Another respondent commented that;

‘Patients require adequate time for history and examination, but clinicians do not provide it due to the high workload caused by the high patient volume compared to the existing diagnosticians. For example, despite the fact that Nakaseke Hospital is supposed to have 12, only five clinicians have been assigned. Otherwise, we will not be able to avoid the next syndrome' (Clinical Officer 1).

The phrase "next syndrome" refers to a practice in the outpatient department in which clinicians quickly attend to patients with less consultation time than the standard fifteen minutes.

Diagnostician’s Experience

Sometimes it is trainees and nurses who diagnose and admit patients when they should only do so under the supervision of senior clinicians or physicians. Because they are less experienced in the diagnostic process, they are more likely to be challenged by rare diseases with which they are unfamiliar. This was emphasized by a medical officer who explained this:

‘Nurses and students are admitting patients despite their inexperience due to hospital understaffing. Such people are more likely to be at fault for misdiagnosis, especially when dealing with unusual cases.' (Medical Officer 2).

This has been attributed to supervisors occasionally failing in their supervisory role, resulting in inexperienced diagnosticians making poor diagnostic decisions.

Failure To Keep Up With Advancements In Disease Diagnosis

Most diseases today require the use of modern investigative technologies that older clinicians may not have encountered during their training or orientation. As a result, clinicians who received their last clinical training more than two decades ago may be inefficient diagnosticians in modern practice. A medical officer explained that;

Some clinicians are accustomed to treating patients symptomatically because diagnostic technologies were scarce at the time of their training. Yes, symptom-based Integrated Management of Childhood Illnesses is common (IMCI) even for adult illnesses. On the other hand, new diagnostic technology and methods, have been introduced for which some clinicians have not received training (Medical Officer 2).

Similar explanations were given for clinicians trained more than 20 years ago who may struggle to diagnose emerging diseases such as SARS, the coronavirus disease of 2019 (COVID-19), and hemorrhagic fevers such as Ebola, which they may have not received any training about.

New diseases are emerging all the time. Now, if someone qualified in 2018 and hasn't kept up with new information, they may be unfamiliar with COVID-19, for example. People who are in training are more likely to be familiar with COVID-19 because they are learning about it (Clinical Officer 3).

New knowledge and advances in the presentation and diagnosis of some common diseases, such as tuberculosis, have been discovered. Unfortunately, the use of such innovations for diagnosis may not be known to cliniians who graduated a long time ago, especially those who have not received continuous professional training in this field. According to one of the participants:

We have Clinical Officers who are about to retire and are incapable of properly diagnosing HIV-related comorbidities. Clinicians who trained in the 1990s, when HIV/AIDS comorbidities such as Cryptococcal meningitis did not exist, or those that were poorly documented, such as Tuberculosis and others, are unable to accurately diagnose such diseases, particularly if they have not received refresher trainings since graduating from medical school (Medical Officer 1).

Another finding was that newly qualified clinicians would have a harder time diagnosing rare diseases that were not common at the time or during their training than experienced clinicians who are knowledgeable about such diseases because they encountered them at the time of their training or early practice.

Service delivery and misdiagnosis

Specialists' Delayed Opinions

According to some participants, the misdiagnosis of certain diseases, such as psychiatric, nutritional, orthopedic or ophthalmic diseases, was caused by the failure to timely seek the opinion of specialized diagnosticians. This challenge is explained in detail in the following excerpt;

‘Patients seeking treatment for psychiatric, ophthalmic, or malnutrition issues are frequently admitted without a diagnosis, while clinicians consult with specialists to ensure the correct diagnosis is made. Unfortunately, psychiatrists, ophthalmic clinical officers, and nutritionists are not always available in the outpatient department (OPD) of the hospital where patients are admitted. To make matters worse, most hospitals only have one specialist on staff (Medical Officer 7).

However, some respondents argue that patient misdiagnosis for conditions such as those mentioned above may be caused by clinicians always waiting for and over-relying on specialist opinions and eventual diagnoses. When a psychiatric condition is suspected, a patient may be misdiagnosed or miss another non-psychiatric diagnosis because the psychiatrists focus solely on mental health without considering other non-psychiatric options.

'... Diagnosis is made by people based on their education and specialization. An admitting psychiatric clinical Officer may examine a patient from the perspective of mental health, and a general Medical Officer may then make a different diagnosis based on a more thorough examination of the patient (Medical Officer 5).

Inappropriate Use Of Diagnostic Guidelines

Some participants believe that clinicians may occasionally misapply diagnostic guidelines when treating patients. For example, some clinicians use the Integrated Management of Childhood Diseases (IMCI) guidelines, which are intended primarily for the diagnosis and treatment of childhood illnesses, to diagnose adult diseases. According to one of the interviewees;

Some clinical officers incorrectly use IMCI guidelines intended for children to guide diagnosis of adult conditions, particularly those presenting with fever, cough, diarrhoea, and difficulty breathing (Medical Officer 1).

The application of IMCI guidelines to pediatric patients in situations where they are inappropriate has also been criticized. As illustrated by the following comment, some participants who shared this concern were particularly critical of clinicians who were trained during the peak of IMCI training and felt indecisive about when to use the guidelines and when not to.

‘Clinical officers who were trained in the past are used to treating patients symptomatically with the IMCI, but many new methods and technologies of disease diagnosis and management have been introduced that they are unaware of' (Medical Officer 6).

Misdiagnosis Is Fairly More Common Among Referred Patients

When compared to those referred from lower-level healthcare facilities or clinics, patients who were not referred had a significantly lower likelihood of misdiagnosis (49%) (aOR = 0.51; 95% CI = 0.31–0.86, p = 0.011). Participants in the qualitative interviews provided explanations for this finding, including the possibility of sending misleading information and incorrect test results with referred patients, as well as referrals from inexperienced healthcare professionals where the diagnosis is not validated by the receiving clinicians.

Misleading Referral Information

According to some qualitative interview participants, patients referred from other healthcare institutions may provide inaccurate information, such as false medical histories or test results. As a result, receiving clinicians may make a mistake by failing to validate the accuracy of the referral information with a repeat medical history and examination:

"... I believe it's because of the information provided by the referred patient. The patient's information may bias a clinician, and they may be persuaded to believe that is the most likely diagnosis and keep it.' (Medical Officer 4).

Less qualified or experienced personnel, such as nursing assistants from lower-level health facilities, may refer patients to higher-level facilities in some cases. In such cases, it is critical for receiving clinicians to review and validate the referral information in order to avoid patient misdiagnosis, which such receiving clinicians may not have done:

Patients at lower-level health facilities are sometimes managed by less qualified staff such as nursing assistants because clinicians are too busy with administrative work. As a result, when receiving clinicians do not double-check the referral information and instead rely on the referral diagnosis, such patients are misdiagnosed. (Clinical Officer 1)

Some clinicians, according to the respondents, believe that a referral diagnosis made by an experienced or more qualified diagnostician at the referral center is correct and does not need to be validated.

‘Some clinicians prefer not to take a new history or examine referred patients if the referral is made by someone more experienced or qualified. They reason and believe that because the patient was referred by a senior person, such a diagnosis must be correct, and they will stick with it even if the referring doctor or senior clinician made a mistake (Clinical Officer 4).

According to the participants, one common cause of incorrect information is when patients bring in test results from private clinics. In particular, clinicians may be more prone to making errors if they heavily rely on the results of sonographic testing or typhoid tests from private clinics.

'... Occasionally, referred patients present with ambiguous laboratory and sonographic results. You get sonographic results from someone who isn't even a sonographer and has only recently learned on the job how to do sonography. As a medical officer reviewing the admitted patient, you become perplexed when you discover insufficient evidence to support the clinician's assigned diagnosis based on investigation reports obtained from private clinics (Medical Officer 5).

Medical technology and equipment and misdiagnosis

According to the participants, placing excessive reliance on laboratory tests for patient diagnosis can result in misdiagnosis. The interviews also highlighted the challenges associated with inadequate medical examination and diagnostic technologies, which can lead to inaccurate diagnoses.

Lack Of Access To Investigations

The majority of participants identified difficulty in accessing diagnostic tests as one of the factors contributing to patient misdiagnosis, particularly when diagnosticians ordered tests, such as the Cell Blood Count (CBC) and X-ray imaging, could not be performed in hospitals due to machine breakdown. As described below, if clinicians only use symptomatic management when laboratory testing would help them make a better diagnosis, they risk making wrong diagnostic decisions.

‘A child may have a fever and a cough, and a chest examination may suggest pneumonia or tuberculosis. You would have preferred to conduct chest X-ray investigations to confirm the diagnosis if the X-ray machine had been operational. As a result, you make a pneumonia diagnosis based on clinical signs and symptoms that is later proven to be incorrect.' (Clinical Officer 3).

Ineffective Diagnostic Tests

Certain laboratory tests, such as the Malaria Rapid Diagnostic Tests (MRDT) and peptic ulcer disease (PUD) rapid diagnostic tests, have been identified by some as possible causes of patient misdiagnosis due to their sensitivity and specificity challenges. For example, while PUD rapid diagnostic tests are highly sensitive, they have low specificity. This means that many patients are initially diagnosed with PUD but are later found to have a different condition after being thoroughly reviewed. One participant provided the following comment:

The poor quality of the rapid diagnostic strips we use contributes to disease misdiagnosis, such as peptic ulcers. Whether or not a patient is experiencing epigastric pain, an H-pylori blood test is always performed, and the results are always positive. It would not be surprising if a child chosen at random from the consultation line tested positive with the RDT despite having no signs or symptoms of peptic ulcers. The same can be said of rapid malaria diagnostic tests' (Clinical Officer 1).

Patients may be misdiagnosed when clinicians rely solely on laboratory tests without conducting thorough medical examinations and taking adequate patient histories. This is often the case with clinicians who depend heavily on investigations and fail to take into account other critical diagnostic factors.

Leadership, governance and organization of care

Inadequate Clinical And Trainee Supervision

The interviewees also indicated that diagnostic practitioners were not being properly supervised during patient care. They observed that in some circumstances clinicians were delegating diagnostic responsibilities to interns without adequate supervision, which resulted in poor diagnostic decisions that were not identified until the patient was further examined in the wards.

Participants also stated that diagnosticians were not adequately supervised while providing care. Clinicians were also observed entrusting diagnostic duties to student trainees on practicum placements without supervising them, resulting in poor diagnostic decisions that went unnoticed until the patient was reviewed on the wards later (Clinical Officer 4).

Misdiagnosis Is Rarely Discussed In Hospitals

The study also found that there was little discussion among diagnosticians about misdiagnosis and diagnostic errors in general. Additionally, senior diagnosticians were not seen to provide feedback on diagnostic errors to their junior colleagues, which could have helped improve diagnostic services.

As people involved in the diagnostic procedure, we don't meet on a regular basis, but we recently discussed misdiagnosis. We have, on occasion, reached an agreement on some of the misdiagnosis and patient management irregularities (Clinical Officer 1).

According to some participants, most discussions about patient safety centered on medication errors.

We have recently primarily discussed drug-related issues. We requested that the pharmacist gather sufficient evidence so that we could discuss drug-related issues. We previously discussed and distributed information to those prescribers, focusing on polypharmacy and adverse drug reactions (Medical Officer 4).

Information, communication and technology

Patients Misguiding Clinicians

Some participants believe that misdiagnosis is caused by patients providing false information to clinicians or manipulating clinicians. Certain cultures believe that unless a patient is given an injection while hospitalized, they have not received proper treatment. As a result, patients may fabricate medical histories to justify the need for an injection and hospitalization:

'... You are aware that the policy forbids you from administering Artesunate intravenously if the malaria is mild.' Because the 'balalo' (cattle keepers) in this area believe that only “kikato”, (injections) can heal, the clinician may be duped into making a severe malaria diagnosis that necessitates injection when it is not required but is done to appease the patient (Medical Officer 1).

Patients who are familiar of their chronic conditions may mislead clinicians by providing skewed medical histories based on their own biases. If the clinician does not recognize the patient's insight and adjusts their diagnosis according to clinical diagnostic standards, this could lead to a misdiagnosis.

The objective of this research was to establish the correlation between various factors such as patient, disease, contextual, and healthcare system factors with patient misdiagnosis in general hospitals located in central Uganda. The study discovered that 9.2% of inpatients were misdiagnosed upon admission to the outpatient clinic and emergency room. Patient age, specifically being elderly, and male gender were associated with misdiagnosis. The research also found that misdiagnosis was related to disease factors, such as treatment for rare diseases, underlying diseases, and the presence of multiple diseases (multimorbidity). The study identified diagnoses made at night as the contextual factor linked with misdiagnosis. Patient referral was identified as a healthcare system factor associated with misdiagnosis.

Patient-related Factors And Misdiagnosis

The findings of this study indicated that the probability of misdiagnosis was greater among elderly patients in comparison to younger ones. The findings validate previous research indicating that healthcare systems face significant deficiencies in offering medical care to the elderly, and that identifying and treating health issues in older patients may be more complex and demanding than in younger ones, thus increasing the chance of incorrect diagnoses (47–49). Elderly patients typically have multiple disorders and potential failures that can weaken other body systems (50). The lack of geriatric care skills among Ugandan clinicians has also complicated diagnostic decision-making in these patients (51). Diagnosticians face challenges when it comes to communicating with senior patients who frequently suffer from various health problems (49). These findings underline the importance of prioritizing elderly patients' health in the Ugandan healthcare system to address the identified factors contributing to suboptimal diagnostic outcomes.

As per the results of this study, male patients had twice the likelihood of being linked to misdiagnosis compared to female patients, which is a rather unprecedented finding. Earlier research had shown that women were more prone to misdiagnosis even if they presented comparable disease symptoms with that of the men (12, 52, 53). This discovery could be attributed to a variety of factors. Historically, medical systems have been primarily concerned with the well-being of women and children, causing healthcare professionals to be less familiar with men's health issues (54). Furthermore, because men are less familiar with healthcare systems than women, they are less active in diagnostic decision making. Because of their limited participation, their discussions with diagnosticians are less diagnostically revealing, resulting in poor diagnostic outcomes (55). Therefore, it is crucial to promote men to seek medical care early and to educate them on how to communicate effectively with diagnosticians. By doing so, men's health behaviors and their participation in the diagnostic decision-making process could be enhanced.

Disease Related Factors And Misdiagnosis

The study findings indicate that patients who have multiple health conditions are about five times more prone to misdiagnosis. These findings are in line with a Japanese study that revealed that patients with several ailments had a 1.2-fold higher risk of diagnostic errors (56). Furthermore, according to Aoki and Watanuki (2020), misdiagnoses occurred in 3.9% of patients without multimorbidity but in 4.9% of multimorbid patients (57). As this study demonstrated, diagnostic decision-making in multimorbid patients can be difficult when physicians are unfamiliar with the cases that multimorbid patients may be having. Clinician's attention can be drawn to the most obvious ailments, especially when the patient has an underlying condition, and unusual signs can be missed. According to the study results, patients with underlying medical conditions were more likely to be misdiagnosed than patients who did not report having any.

According to Hausmann et al. (2019), the presence of underlying diseases may increase the possibility of diagnostic decision bias (es). In a way, when a patient mentions an underlying disease, diagnosticians can think along those lines, rather than thinking outside the box of underlying diseases. If the underlying condition is not mentioned, the clinician may overlook it, even when it is the most likely pathology (58).

Contextual Factors And Misdiagnosis

It has been found that the night shift is the most common time of day for patient misdiagnosis. Other studies generally support the finding that night shift work is associated with diagnostic errors, with drowsiness, sleep deprivation, fatigue, anxiety, and attention deficit contributing to poor cognitive performance and errors (59–62). Misdiagnosis at night has been attributed to a lack of access to examination and diagnostic equipment and tools, such as thermometers, blood pressure monitors, and laboratory tests and investigations. Additionally, when faced with a diagnostic problem late at night, clinicians do not have access to more experienced or qualified staff.

Another element that can impede a precise diagnosis is the absence of laboratory and diagnostic examinations during night shifts. Night shifts typically involve attending to more severe cases, often referred from other hospitals with acute conditions that need immediate attention (63). According to the findings presented in this study, it is advised to prioritize night shifts to improve diagnostic accuracy. Healthcare administrators can execute night management initiatives that involve proper scheduling, staffing, coordination, and the provision of essential resources to facilitate accurate patient diagnosis during night shifts.

The study results revealed no significant correlation between misdiagnosis and admission day (weekdays or weekends), indicating that misdiagnosis was not affected by the day of the week. This emphasizes the importance of being vigilant in accurately diagnosing patients and implementing measures to improve diagnostic procedures throughout the week. Nonetheless, previous research has found that patients who visit hospitals on weekends are more likely to be critically ill (64). Weekend shift care is generally considered challenging due to staffing and logistical difficulties, resulting in a higher risk of misdiagnosis (65). Consequently, if this scenario was the case in the hospitals under investigation, those patients may have been among those referred to other healthcare facilities for further treatment.

Patients who were not referred from lower-level primary care settings had a lower risk of being misdiagnosed. Misdiagnosis of referred patients has been associated with cognitive biases, in which treating physicians are more likely to accept the referring diagnostician's diagnosis without question (20). When the referring physician is incorrect and the receiving physician fails to investigate further but continues to use the same diagnosis, misdiagnosis can occur. Moreover, a study on diagnostic errors among tuberculosis patients discovered that misdiagnosis frequently occurred among referred patients due to uncoordinated referral systems, in which the referring physician offered little information to the receiving physicians concerning the reasons for the referral (66). In addition, referred cases are more likely to be complicated than non-referred cases, making diagnosis more difficult (67). If special attention is paid to referred cases during the diagnostic process, the outcome of the diagnosis in these patient groups is likely to improve.

Health System Factors And Misdiagnosis

This study has identified several healthcare system factors that lead to patient misdiagnosis. Service delivery problems, such as poor referral systems, flawed investigations, and delayed referrals, are examples. In addition, misdiagnosis can result from over-reliance on investigations, which reduces clinicians' cognitive abilities during the diagnostic process. The training of experienced clinicians appears to be outdated, as they may not be familiar with current medical technologies and new disease states not covered in previous training. Conversely, newly trained and inexperienced clinicians may misdiagnose rare diseases due to a lack of exposure to them (68). Although healthcare system design does not facilitate workplace mentoring, reporting errors, and providing feedback on errors, including diagnostic errors, diagnosticians should prioritize misdiagnosis and diagnostic error prevention in academic and continuing medical education by enhancing everyday clinical practice and awareness of healthcare system factors linked to diagnostic errors.

Implication Of The Findings And Future Direction

The study's findings, particularly the factors that contribute to diagnostic errors, should be used to educate healthcare providers at all levels. Clinicians should be trained to recognize, diagnose, and treat patients who have been identified in this study as having risk factors for misdiagnosis. The diagnostician's ability to adapt to changing contexts in patient disease profiles, such as increased life expectancy, new diseases associated with aging, new technologies, rare diseases, increased workload, comorbidities, and underlying diseases, as well as improved care involving more patient queries, should be strengthened. To avoid catching clinicians off guard, both human (advisory) and technical support should be available at all times, particularly during investigations.

Study Limitations

The study's findings have three limitations to consider. First, due to the paper-based nature of the records, many of them contained errors, which resulted in their exclusion and may have impacted the study results. Second, as the study only focused on general hospitals, the findings may not be entirely representative of higher levels of the healthcare system, such as regional and national hospitals, where disease profiles may differ. Third, there was a lack of data on diagnosticians' variables, such as age, experience, gender, and training, which prevented the determination of their association with patient misdiagnosis. Hence, it is recommended to conduct research at higher levels of the healthcare system, taking into account diagnosticians' variables.

To avoid misdiagnosing patients, it is necessary to take into account factors such as the healthcare system, the disease, the patient, and the operational context. Those responsible for healthcare systems should create processes to promptly and methodically identify cases of misdiagnosis or diagnostic errors, as well as establish a feedback mechanism to develop strategies that can improve diagnostic results. Hospitals should provide diagnostic error training, specifically for resident clinicians and interns, and allocate the necessary resources to improve diagnostic outcomes.

Ethics approval and consent to participate: This study was approved by the Department of Health Policy, Planning, and Management (HPPM) at the University of Ghana, the Mildmay Uganda Research Ethics Committee (MUREC), and the Uganda National Council of Science and Technology under reference numbers HP/AC.12 /1 /2017, #REC REF 0505 2020) and HS826ES. The District Health Officers (DHOs) and Medical Superintendents of the districts where the hospitals are located gave us permission to conduct the studies in their respective district hospitals. We asked MUREC to waive the requirement for patients' consent to access their records because, after evaluating the study's objectives, we determined that obtaining their information without their knowledge would not harm them. Furthermore, it was not possible to locate all of the patients whose records were included in the study and obtain their permission to review their records. All study procedures were conducted in accordance with the Declaration of Helsinki, UNCST, MUREC and Department of Health Policy, Planning and Management, University of Ghana. Informed consent to participate in the study was obtained from all of the participants in the qualitative part of the study

Consent for publication: Not Applicable.

Availability of data and materials: All the necessary data has been provided in here

Competing interests: The authors declare no competing interests

Funding: There was no funding received for this study

Authors' contributions: All authors made significant contributions to the conception, study design, implementation, data collection, analysis, and interpretation of the reported work. They also participated in drafting, revising, or critically reviewing the article. They all gave final approval of the manuscript and agree to be responsible for all aspects of the work.

Acknowledgements: We thank the medical superintendents who gave permission to conduct the study in their hospitals. We also acknowledge the archivists and hospital staff involved in this study by retrieving the medical records and those who participated as data collectors.

Schiff GD, Hasan O, Kim S, Abrams R, Cosby K, Lambert BL et al. Diagnostic Error in Medicine. Vol. 169, Archives of Internal Medicine. 2009.
Mahumud RA, Sultana M, Sheikh N, MN A, Mitra DK, Sarker AR. Diagnostic Errors in Low and Middle-Income Countries : Future Health and Economic Burden for Patient Safety. Austin Emerg Med. 2016;2(8):8–10.
Zwaan L, Singh H. The challenges in defining and measuring diagnostic error. Diagnosis. 2015;2(2):97–103.
Johnson CC, Fonner V, Sands A, Ford N, Obermeyer CM, Tsui S et al. To err is human, to correct is public health: a systematic review examining poor quality testing and misdiagnosis of HIV status. J Int AIDS Soc [Internet]. 2017;20(Suppl 6):21755. Available from: https://doi.org/10.7448/IAS.20.7.21755
Masamba LPL, Jere Y, Brown ERS, Gorman DR. Tuberculosis Diagnosis Delaying Treatment of Cancer: Experience From a New Oncology Unit in Blantyre, Malawi. J Glob Oncol. 2016;2(1):26–9.
Buyego P, Nakiyingi L, Ddungu H, Walimbwa S, Nalwanga D, Reynolds SJ, et al. Possible misdiagnosis of HIV associated lymphoma as tuberculosis among patients attending Uganda Cancer Institute. AIDS Res Ther. 2017;14(1):10–5.
Ghai RR, Thurber MI, El Bakry A, Chapman CA, Goldberg TL. Multi-method assessment of patients with febrile illness reveals over-diagnosis of malaria in rural Uganda. Malar J. 2016;15(1):1–6.
Croskerry P, Cosby KS, Graber ML, Singh H. Diagnosis: Interpreting the shadows. London: CRC Press Taylor and Francis Group; 2017.
GUO S, ROUDSARI A, GARCEZ A, d’Avila. Modelling clinical diagnostic errors : A system dynamics approach. Driving Quality in Informatics:Fulfilling the Promise. Victoria; 2015.
Graber ML, Franklin N, Gordon R. Diagnostic error in internal medicine. Arch Intern Med. 2005;165(13):1493–9.
World Health Organization. Summarry of the evidence on patient safety: Implication for research [Internet]. World Alliance for Patient Safety. Geneva. ; 2008. Available from: http://www.scielo.org.co/pdf/inf/v14n4/v14n4a08.pdf
NAM. Improving diagnosis in health care. National A. Balogh EP, Miller BT, Ball JR, editors. Improving Diagnosis in Health Care. Washington DC: The National Academies Press; 2015.
Hussain F, Cooper A, Carson-stevens A, Donaldson L, Hibbert P, Hughes T, et al. Diagnostic error in the emergency department : learning from national patient safety incident report analysis. BMC Emerg Med. 2019;19(1):1–9.
Katongole SP, Akweongo P, Anguyo RD, Kasozi DE, Adoma-Afari A. Prevalence and classification of misdiagnosis among hospitalised patients in five general hospitals of Central Uganda. Clin Audit. 2022;14:65–77.
Mukasa N. Uganda Healthcare system profile: Background, Organization, Polices and Challenges. J Sustain Reg Heal Syst [Internet]. 2012;1(1):2–10. Available from: http://projektai.vu.lt/erasmusmundus/sites/default/files/u3/Regional Health Systems_Issue 1_Volume 1_May_2012.pdf
Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373–9.
Kedar I, Ternullo JL, Weinrib CE, Kelleher KM. Brandling-bennett H. Information in practice. BMJ. 2003;326:696–9.
Epstein HM. Why your child is more likely to be misdiagnosed than you are [Internet]. Dx IQ No. 3: Be informed and prepared for every medical interaction. 2019 [cited 2019 Sep 11]. Available from: https://www.improvediagnosis.org/dxiq-column/why-your-child-is-more-likely-to-be-misdiagnosed-than-you-are/
Mohamed F, Mamede S, Mohamedani M, Al I. The effect of patients ’ appearance on doctors ’ diagnostic decision making: Do poor people get poorer medical care ? Heal Prof Educ [Internet]. 2016;2(1):18–23. Available from: http://dx.doi.org/10.1016/j.hpe.2016.01.011
Lighthall GK, Vazquez-Guillamet C. Understanding decision making in critical care. Clin Med Res. 2015;13(3–4):156–68.
Kaisey M, Solomon AJ, Luu M, Giesser BS, Sicotte NL. Incidence of multiple sclerosis misdiagnosis in referrals to two academic centers. Mult Scler Relat Disord. 2019;30:51–6.
Newman-Toker DE, Schaffer AC, Yu-Moe CW, Nassery N, Saber Tehrani AS, Clemens GD, et al. Serious misdiagnosis-related harms in malpractice claims: The “Big Three” – vascular events, infections, and cancers. Diagnosis. 2019;6(3):227–40.
Marquez PV, Farrington JL. “The challenge of non-communicable diseases and road traffic injuries in Sub-Saharan Africa. An Overview [Internet]. Washington, DC: The world bank. Washington, DC; 2013. Report No.: CC BY-NC-ND 3.0 IGO. Available from: https://openknowledge.worldbank.org/handle/10986/16451
Kostopoulou O, Nath R, Delaney B. Predictors of Diagnostic Accuracy and Safe Management in Difficult Diagnostic Problems in Family Medicine. Med Decis Mak. 2008;28:668–80.
Maude J. Differential diagnosis: the key to reducing diagnosis error, measuring diagnosis and a mechanism to reduce healthcare costs. Diagnosis. 2014;1(1):107–9.
Kijima T, Akai K, Nabika T, Taniguchi E, Matsushita A, Ishibashi Y. Diagnostic Difficulties and Factors Affecting Diagnosis in Acutely Ill Elderly Japanese Patients Living at Home. Int J Gerontol [Internet]. 2018;12(4):326–30. Available from: https://doi.org/10.1016/j.ijge.2018.05.004
Abe T, Tokuda Y, Shiraishi A, Fujishima S, Mayumi T, Sugiyama T, et al. In-hospital mortality associated with the misdiagnosis or unidentified site of infection at admission. Crit Care. 2019;23(1):1–9.
Hautz SC, Schuler L, Kämmer JE, Schauber SK, Ricklin ME, Sauter TC, et al. Factors predicting a change in diagnosis in patients hospitalised through the emergency room: a prospective observational study. BMJ Open. 2016;6(5):e011585.
Fatima S, Shamim S, Butt AS, Awan S, Riffat S, Tariq M. The discrepancy between admission and discharge diagnoses : Underlying factors and potential clinical outcomes in a low socioeconomic country. PLoS One [Internet]. 2021;16(6):e0253316. Available from: http://dx.doi.org/10.1371/journal.pone.0253316
Avelino-Silva JT, Alan SM. Diagnostic discrepancies between emergency department admissions and hospital discharges among older adults: Secondary analysis on a population-based survey. Sao Paulo Med J. 2020;138(5):359–67.
World Health Organization (WHO). WHO-FIC Content Model Reference Guide [Internet]. Geneva. ; 2021. Available from: https://icd.who.int/browse11.
Harel Z, Silver SA, Mcquillan RF, Weizman AV, Thomas A, Chertow GM, et al. How to Diagnose Solutions to a Quality of Care Problem. Clin J Am Soc Nephrol. 2016;11(5):901–7.
Berenson RA, Upadhyay DK, Kaye DR. Placing Diagnosis Errors on the Policy Agenda. Washington, DC: Robert Wood Johnson Foundation; 2014.
Finlay L. Engaging Phenomenological Analysis. Qual Res Psychol [Internet]. 2014;11(2):121–41. Available from: http://dx.doi.org/10.1080/14780887.2013.807899
Guest G, Bunce A, Johnson L. How Many Interviews Are Enough?: An Experiment with Data Saturation and Variability. Field methods. 2006;18(1):59–82.
The Republic of Uganda. Scheme of service for the medical clinical officers cadre. Kampala: Ministry of Health; 2021.
Matsiko CW, Kiwanuka J. A review of human resourcesfor health in Uganda. Heal Policy Dev. 2003;1(1):15–20.
The Republic of Uganda. Job descriptions and specifications for jobs in local government. Kampala; 2011.
Kohn MA. Understanding evidence-based diagnosis. Diagnosis. 2014;1(1):39–42.
Richards HM, Schwartz LJ. Ethics of qualitative research: are there special issues for health services research? Fam Pract [Internet]. 2002;19(2):135–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11906977
Chan ZC, Fung Y, Chien W. Bracketing in phenomenology : Only undertaken in the data collection and analysis process ? Qual Rep. 2013;18(59):1–9.
Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol [Internet]. 2006;3(2):77–101. Available from: http://eprints.uwe.ac.uk/11735
Creswell JW. Research Design: Qualitative, Quantitative and Mixed Method Aproaches. In: SAGE Publications [Internet]. 2007. p. 203–23. Available from: http://libproxy.unm.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=a9h&AN=51827937&site=eds-live&scope=site%5Cnhttp://content.ebscohost.com.libproxy.unm.edu/ContentServer.asp?T=P&P=AN&K=51827937&S=R&D=a9h&EbscoContent=dGJyMNLr40SeprI4
Creswell JW, Poth CN. In: Creswell JW, Poth CN, editors. Qualitative inquiry and research design: Choosing among five approaches. 4 ed. Sage publications; 2018.
Tracy SJ, Sarah J, Tracy. Qualitative Research Methods: Collecting Evidence, Crafting Analysis, Communicating Impact [Internet]. Edition 1st, editor. Vol. 43, Revija za sociologiju. Chichester, West Sussex: Blackwell Publishing; 2013. 1–341 p. Available from: http://hrcak.srce.hr/index.php?show=clanak&id_clanak_jezik=162194
Creswell JW. Qualitative inquiry and research design: Choosing among five approaches. Volume 2, 2nd ed. London, California and New Delhi: SAGE Publications; 2007. pp. 1–395.
Hajjaj FM, Salek MS, Basra MKA, Finlay AY. Non-clinical influences on clinical decision-making: a major challenge to evidence-based practice. J R Soc Med [Internet]. 2010;103(5):178–87. Available from: http://journals.sagepub.com/doi/10.1258/jrsm.2010.100104
World Health Organization. Delivering quality health services [Internet]. World Health Organization, World Bank Group, OECD. Geneva. ; 2018. Available from: http://apps.who.int/bookorders.
Maniragaba F, Nzabona A, Asiimwe JB, Bizimungu E, Mushomi J, Ntozi J, et al. Factors associated with older persons ’ physical health in rural Uganda. PLoS ONE. 2019;14(1):e0209262.
Singh I. Assessment and Management of Older People in the Assessment General Hospital Setting. Challenges older care [Internet]. 2016;37. Available from: http://dx.doi.org/10.5772/64294
Ssensamba JT, Mukuru M, Nakafeero M, Ssenyonga R, Kiwanuka SN. Health systems readiness to provide geriatric friendly care services in Uganda: a cross-sectional study. BMC Geriatr. 2019;19(256):1–13.
NAM. In: Balogh EP, Miller BT, Ball JR, editors. Improving Diagnosis in Health Care. 1 ed. Washington, DC: The National Academies Press; 2015. p. 472.
Newman-Toker DE, Moy E, Valente E, Coffey R, Hines AL. Missed diagnosis of stroke in the emergency department: a cross-sectional analysis of a large population-based sample. Diagnosis. 2014;1(2):155–66.
Nuzzo J. Bias against men’s issues within the United Nations and the World Health Organization: A content analysis. Psychreg J Psychol [Internet]. 2020;120–50. Available from: https://www.pjp.psychreg.org/wp-content/uploads/2020/12/nuzzo-120-150.pdf
Levinson W, Kao A, Kuby A, Thisted RA. Not all patients want to participate in decision making. J Gen Intern Med. 2005;20(6):531–5.
Panagioti M, Stokes J, Esmail A, Coventry P, Cheraghi-Sohi S, Alam R, et al. Multimorbidity and patient safety incidents in primary care: A systematic review and meta-analysis. PLoS ONE. 2015;10(8):1–30.
Aoki T, Watanuki S. Multimorbidity and patient- reported diagnostic errors in the primary care setting: multicentre cross- sectional study in Japan. BMJ Open. 2020;10(8):10e039040.
Hausmann D, Kiesel V, Zimmerli L, Schlatter N, von Gunten A, Wattinger N, et al. Sensitivity for multimorbidity: The role of diagnostic uncertainty of physicians when evaluating multimorbid video case-based vignettes. PLoS ONE. 2019;14(4):1–16.
Beglinger B, Rohacek M, Ackermann S, Hertwig R, Karakoumis-Ilsemann J, Boutellier S, et al. Physician’s first clinical impression of emergency department patients with nonspecific complaints is associated with morbidity and mortality. Med (Baltim). 2015;94(7):e374.
Hughes V. Is there a relationship between night shift and errors ? What nurse leaders need to know. Athens J Heal. 2015;3(3):217–28.
Lazzari C, Shoka A, Mousailidis G. Sleep deprivation in healthcare professionals and medical errors : how to recognize them ? Sleep Med Disord Int J. 2018;2(1):15–6.
Hamilton-fairley D, Coakley J, Moss F. Hospital at night: an organizational design that provides safer care at night. BMC Med Educ. 2014;14(Suppl 1):17.
Mathew A, Fyyaz SA, Carter PR, Potluri R. The enigma of the weekend effect. J Thorac Dis. 2018;10(1):102–5.
Sears D, Mpimbaza A, Kigozi R, Sserwanga A, Chang MA, Kapella BK, et al. Quality of inpatient pediatric case management for four leading causes of Child Mortality at Six Government-Run Ugandan Hospitals. PLoS Med. 2015;10(5):e0127192.
Neshati H, Sheybani F, Naderi H, Sarvghad M, Soltani AK, Efterkharpoor E et al. Diagnostic errors in Tuberculous patients: A multicenter study from a developing country. J Environ Public Health. 2018;2018.
Bracho-Blanchet E, Cazares-Rangel J, Zalles-Vidal C, Davila-Perez R. Misdiagnosis and quality of management in paediatric surgical patients referred to a tertiary care hospital. J Clin Diagnostic Res. 2014;8(4):1–5.
Vandeborne L, Overbeeke E, Van, Dooms M, Beleyr B, De, Huys I. Information needs of physicians regarding the diagnosis of rare diseases : a questionnaire-based study in Belgium. Orphanet J Rare Dis. 2019;14(1):1–11.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Factors associated with misdiagnosis of hospitalised patients: A mixed method study in five general hospitals of Central Uganda

Status:

Version 1

Abstract

Background

Methods

Study design

Setting

Quantitative Data

Inclusion And Exclusion Criteria

Data Collection

The Secondary Outcomes:

Qualitative Study

Qualitative Data Analysis

Integration Of Mixed Methods

Results

Characteristics of patient records reviewed

Participants For The Qualitative Analysis

Primary Outcomes

Factors Associated With Misdiagnosis At Bivariate Analysis

Diseases/conditions Involved In Misdiagnosis

The contextual factors associated with misdiagnosis

Patients’ Misdiagnosis And Night Admission

Clinicians Fatigued At Night

Difficult To Do Investigations And Examination At Night

Patient Management At Night Is Centred On Stabilizing The Patient

Difficult To Consult Colleagues At Night When Challenged

Patient related factors and misdiagnosis

Male Patients Are More Likely To Be Misdiagnosed

Patients’ Age

Children Less Likely To Falsify Illnesses

Fewer Pathologies In Children Means Fewer Incorrect Diagnoses

Complexities Of Adult Conditions

Disease Related Factors And Misdiagnosis (Dup: Abstract ?)

Presence of underlying diseases and misdiagnosis

Multimorbidity And Misdiagnosis

Rare Or Uncommon Diseases

Health system factors associated with misdiagnosis

Human resources for health and misdiagnosis

Scarcity Of Human Resources

Diagnostician’s Experience

Failure To Keep Up With Advancements In Disease Diagnosis

Service delivery and misdiagnosis

Specialists' Delayed Opinions

Inappropriate Use Of Diagnostic Guidelines

Misdiagnosis Is Fairly More Common Among Referred Patients

Misleading Referral Information

Medical technology and equipment and misdiagnosis

Lack Of Access To Investigations

Ineffective Diagnostic Tests

Leadership, governance and organization of care

Inadequate Clinical And Trainee Supervision

Misdiagnosis Is Rarely Discussed In Hospitals

Information, communication and technology

Patients Misguiding Clinicians

Discussion

Patient-related Factors And Misdiagnosis

Disease Related Factors And Misdiagnosis

Contextual Factors And Misdiagnosis

Health System Factors And Misdiagnosis

Implication Of The Findings And Future Direction

Study Limitations

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1