TyphoDx-BD: A Clinically Validated Dataset for Machine Learning-Based Typhoid and Rickettsial Disease Prediction in Bangladesh

This dataset contains detailed clinical records of 1,106 patients collected from Bangladesh between October 27, 2024 and January 2, 2025. Typhoid fever is a serious bacterial infection caused by Salmonella Typhi, which is commonly spread through contaminated food and water. It remains a widespread health issue in many parts of Bangladesh, especially where sanitation and clean water access are limited. The dataset includes results from Widal tests, which are commonly used to detect typhoid and paratyphoid fever. These include antigens such as TO and TH (specific to Salmonella Typhi) and AH and BH (specific to Salmonella Paratyphi). Additionally, the dataset contains values from the Weil-Felix test, including OXK, OX2, and OX19, which are used to help identify rickettsial infections — a group of bacterial infections often presenting with similar symptoms. This dataset has been validated by a certified medical professional and is suitable for use in epidemiological research, diagnostic modeling, and public health analysis.

1. Weil-Felix Test (Rickettsial Infections): OXK, OX2, OX19
→ Antigens used to detect antibodies against rickettsial infections (e.g., scrub typhus, epidemic typhus).
Note: Weil-Felix is nonspecific and largely replaced by serological/PCR tests in modern labs.

2. Widal Test (Typhoid/Paratyphoid Fever): TO (Typhi O), TH (Typhi H), AH (Paratyphi A H), BH (Paratyphi B H)
→ Antigens measuring antibody titers for:
- Salmonella Typhi (TO, TH)
- Salmonella Paratyphi A/B (AH, BH).

3. Typhoid Status:
Typhoid
→ Overall diagnosis (e.g., Minimal/Negative/Positive).

Acute_typhoid (Yes/No)
→ Confirms active infection (likely based on clinical symptoms + high titers).

Paratyphoid_A/B (Yes/No)
→ Specific to Salmonella Paratyphi A or B infections.

4. Rickettsial Suspicion
Rickettsia_Suspect (Yes/No)
→ Flagged if Weil-Felix titers (OXK/OX2/OX19) or symptoms suggest rickettsiosis.

5. Additional Diagnostic Markers
M, A (Binary: 0/1)
→ Likely represent: M: IgM antibody presence (acute phase); A: Agglutination result (if qualitative).

6. Demographic & Privacy Fields: Gender (Male/Female); Age (Years); Encrypted_Name: → Deidentified patient records for ethical compliance.

Dataset Validation:
Validated by: Dr. Prio Gopal Biswas (Medical Doctor)
Hospital: Saranjghola Upazila Health Complex, Bagerhat, Bangladesh
Data Collection Period: October 27, 2024 – January 2, 2025

Ethical Protocols and Statement:
All procedures were conducted in alignment with the ethical guidelines of Daffodil International University (DIU) and in accordance with relevant national and institutional regulations. The Research Ethics Committee (REC) of the Faculty of Science and Information Technology (FSIT) at DIU granted ethical approval for this study under the approval number REC-FSIT-2024-09-17, following a thorough review process. Written informed consent was obtained from all participants.