Page 129 - AIH-1-2

Artificial Intelligence in Health SDoH in clinical narratives

Table 2. Average SDoH mentions from clinical case reports 95% CI: 1.97 – 3.51), gynecology (AOR: 2.45; 95% CI:
(n=463546) between 1975 and 2022 2.01 – 2.99), rehabilitation (AOR: 2.39; 95% CI: 1.31 –
4.35), and forensic medicine (AOR: 2.04; 95% CI: 1.32 –
SDoH SDoH mentions (95% CI) 3.17). Conversely, nephrology (AOR: 0.46; 95% CI: 0.25
Race/ethnicity 383.24 (377.71–388.77) – 0.79) and traumatology (AOR: 0.46; 95% CI: 0.26 – 0.79)
Marital status 42.95 (41.06–44.83) displayed a pronounced negative correlation with mentions
Sexual orientation 11.30 (10.34–12.27) of marital status. Clinical cases pertaining to mental
Immigrant status 6.13 (5.41–6.84) disorders (AOR: 2.14; 95% CI: 1.72 – 2.66) and urogenital
Spiritual beliefs 1.36 (1.02–1.69) diseases (AOR: 1.68; 95% CI: 1.47 – 1.91) were robustly
Homelessness 1.29 (0.97–1.62) associated with mentions of marital status. Authors from
Abbreviations: CI: Confidence interval; SDoH: Social determinants of sub-Saharan Africa also exhibited a marked inclination to
health. mention marital status (AOR: 1.98; 95% CI: 1.32 – 2.96).
Several other factors had associations with the
The specialty of the journal significantly influenced likelihood of mentioning marital status, although more
the likelihood of race/ethnicity mentions. Case reports moderately. Clinical cases covering a broad spectrum
in general medicine were the most likely to include race/ of conditions, such as wounds, neoplasms, infections,
ethnicity (AOR: 2.18; 95% CI: 2.08 – 2.29), followed by digestive, hematological, skin, respiratory, metabolic,
laboratory medicine (AOR: 2.10; 95% CI: 1.94 – 2.28), musculoskeletal, and nervous diseases, as well as those
dentistry (AOR: 1.82; 95% CI: 1.55 – 2.13), and psychiatry related to unspecific signs and symptoms, were linked with
(AOR: 1.82; 95% CI: 1.56 – 2.13). A moderate tendency to slightly fewer mentions of marital status (AOR: 0.51 – 0.77).
mention race/ethnicity was also observed in other journal Journals focusing on gastroenterology and general surgery
specialties (AOR: 1.37 – 1.97) (Table S1). Surgical specialties (AOR: 0.53 – 0.74) also demonstrated a subtle association
were generally less likely to mention race/ethnicity. These with reduced mentions of marital status.
included anesthesiology (AOR: 0.27; 95% CI: 0.20 – 0.37), Lastly, case reports published in the Indian subcontinent
urology (AOR: 0.48; 95% CI: 0.40 – 0.56), traumatology or authored by individuals from the Middle East, the Indian
(AOR: 0.59; 95% CI: 0.50 – 0.70), and general surgery (AOR: subcontinent, North Africa, and Southeast Asia were more
0.61; 95% CI: 0.57 – 0.65). Rehabilitation (AOR: 0.31; 95% CI: inclined to mention marital status (AOR: 1.31 – 1.75).
0.18 – 0.54) and radiology (AOR: 0.40; 95% CI: 0.35 – 0.47) Further details on marital status mentions can be found in
displayed a strong tendency against reporting race/ethnicity Figure 3 and Table S2.
in their clinical cases. Some journal specialties, namely
cardiology (AOR: 0.63; 95% CI: 0.56 – 0.72), pneumology 3.4.3. Sexual orientation
(AOR: 0.75; 95% CI: 0.61 – 0.92), and neurology (AOR: 0.79; The mention of sexual orientation was profoundly
95% CI: 0.72 – 0.87), were slightly less inclined to include
this information in their clinical case reports. correlated with the diagnosis of infectious diseases (AOR:
25.00; 95% CI: 19.68 – 31.75). Other robustly associated
Finally, the primary diagnosis of the clinical case factors include case reports published in South America
was also correlated with the likelihood of race/ethnicity (AOR: 4.04; 95% CI: 1.92 – 8.50) and North America
mentions, although less strongly than the other variables. (AOR: 2.15; 95% CI: 1.31 – 3.55). In contrast, journal
Hematological, eye, stomatognathic, metabolic, skin specialties, such as pediatrics (AOR: 0.16; 95% CI: 0.07
diseases, and infections were significantly associated – 0.39) and surgery (AOR: 0.46; 95% CI: 0.30 – 0.69),
with slightly higher mentions of race/ethnicity (AOR: demonstrated a strong negative correlation with mentions
1.20 – 1.32). Conversely, occupational diseases, wounds and of sexual orientation. A similar trend was also observed
injuries, cardiovascular diseases, nervous system diseases, across a variety of diagnoses, including cardiovascular,
respiratory diseases, and digestive diseases were associated musculoskeletal, and respiratory (AOR: 0.26 – 0.37).
with fewer race/ethnicity mentions (AOR: 0.64 – 0.91).
Authors from the Indian subcontinent (AOR: 0.16; 95%
Detailed information about the AOR of each factor CI: 0.05 – 0.51) and the Middle East (AOR: 0.16; 95% CI:
associated with race/ethnicity mentions can be found in 0.04 – 0.63) were considerably less inclined to mention
Figure 2 and Table S1. sexual orientation. Conversely, authors from North
America (AOR: 1.47; 95% CI: 1.13 – 1.91) and Western
3.4.2. Marital status Europe (AOR: 1.46; 95% CI: 1.15 – 1.87) were more
Mentions of marital status were notably correlated with inclined to mention sexual orientation more frequently
several journal specialties such as psychiatry (AOR: 2.6; than the authors from other regions. Further details on

Volume 1 Issue 2 (2024) 123 doi: 10.36922/aih.2737