Vol. 26 - Num. 104
Original Papers
Is artificial intelligence able to discriminate emergencies?
Raquel Bernal Calmarzaa, Ana Valer Martínezb, María Celada Suárezc, Sara Calmarza Delgadod, Elena Calmarza Delgadod
aPediatra. CS Quince de Mayo. Madrid. España.
bMédico de familia. CS Tarazona. Tarazona. Zaragoza. España.
cMIR-Medicina de Familia. Hospital Universitario Miguel Servet. Zaragoza. España.
dEnfermera. Hospital Ernest Lluch Martin. Calatayud. Zaragoza. España.
Correspondence: R Bernal. E-mail: raquel3433@gmail.com
Reference of this article: Bernal Calmarza R, Valer Martínez A, Celada Suárez M, Calmarza Delgado S, Calmarza Delgado E. Is artificial intelligence able to discriminate emergencies? . Rev Pediatr Aten Primaria. 2024;26:351-60. https://doi.org/10.60147/dce30dee
Published in Internet: 31-10-2024 - Visits: 1793
Abstract
Introduction: in paediatrics, high-frequency emergency department use is defined as repeated emergency visits for reasons that do not require urgent attention or could be managed at a different level of care. Several factors may be associated with this phenomenon, such as socioeconomic, cultural or psychological factors. Its impact on the health care system is significant. Artificial intelligence (AI) has the potential of reducing high-frequency use.
Methodology: we assessed the agreement between the information for 101 diseases common in children provided by Gemini AI, a free and open-access service, and the current scientific evidence. We used the adjusted kappa coefficient in this analysis.
Results: the AI provided responses for all of the 101 diseases considered in the analysis. The kappa coefficient was 0.857 (95% CI, 0.002) for the identification of the disease, 0.888 (95% CI, 0.003) for the identification of warning signs, 0.876 (95% CI, 0.005) for establishing the need to visit the emergency department and 0.915 (95% CI, 0.003) for the appropriate recommendation of measures to be taken.
Conclusions: the text-based artificial intelligence exhibited substantial agreement with protocols used for identification of diseases based on symptoms, and near-perfect agreement for determining the need to visit the emergency department, identifying warning signs and providing therapeutic recommendations. The level of agreement was higher for common diseases and children aged more than 3 months.
Keywords
● Artificial intelligence ● Diagnosis ● EmergenciesINTRODUCTION
High-frequency use of health care services is a growing problem in paediatrics that can have a deleterious impact on care quality and available health care resources. In the emergency care setting, high-frequency use is defined as making recurrent visits to the emergency department for reasons that do not require urgent care or could be managed in a different setting or level of care.
The factors that contribute to high-frequency use of paediatric emergency care services are varied and may include socioeconomic, cultural and psychological factors.1 Possible socioeconomic factors include lack of access to primary care, poverty or immigrant status. Among cultural factors, specific beliefs and practices may contribute significantly to the excessive use of emergency care. The psychological factors include anxiety, stress and lack of trust in primary care services.
High-frequency users have a significant impact on the health care system. They are a small proportion of the paediatric population, but give rise to a large volume of visits.2 This can overburden primary and emergency care services and increase health care costs.
In paediatric care, high-frequency use is a complex problem that needs to be addressed from a multidisciplinary approach. Possible strategies include education of patients and parents, improving access to primary care and the development of specific programmes for the management of high-frequency users.
Artificial intelligence (AI) is revolutionising medicine, with applications in multiple areas, ranging from diagnosis and treatment of diseases to research and development of new drugs.3 Several studies have shown that AI has the potential to be an effective tool to reduce high-frequency use.4 AI can be used to assist patient triage according to the level of urgency and to provide information and support to patients and their parents.5
Although there is evidence that supports the usefulness of AI in assisting triage in paediatric emergency departments, there is no evidence on its performance in supporting parental decision-making regarding the need to visit the emergency department.6
The aim of our study was to measure the level of agreement between the results obtained through Gemini artificial intelligence (formerly known as Bard) and the best available evidence for 101 common and significant diseases in terms of four variables: recognition of the disease, identification of warning signs, need to visit emergency department and measures to be taken at home.
One of the secondary objectives was to assess the agreement between the results yielded by the Gemini AI and the best available evidence for 101 common and significant diseases in terms of the same four variables (recognition of the disease, correction of warning signs, need to visit emergency department and measures to be taken at home) for each disease category. We also aimed to identify factors that could affect the precision and accuracy of diagnosis in the Gemini AI.
MATERIAL AND METHODS
We conducted an observational study comparing the results obtained through prompts issued to the Gemini AI model with the best available scientific evidence, defined as the protocols of the Sociedad Española de Urgencias Pediátricas (SEUP, Spanish Society of Paediatric Emergency Medicine)7 and the algorithms of the Asociación Española de Pediatría de Atención Primaria (AEPap, Spanish Association of Primary Care Paediatrics).8
We used the Gemini artificial intelligence, a machine-learning model trained on a massive dataset of text and code, which has the ability to generate text in response to prompts to providing informational answers in Spanish.
Gemini is one of the world’s largest language models, with 137 billion parameters. This allows it to learn complex language patterns and relationships. Furthermore, it has access to a massive dataset of text and code, allowing it to learn about a broad range of topics. It can generate high-quality, grammatically correct and coherent text and provide informative answers to questions, even if they are open-ended. It is an open-access language model that is free for users, so it is easy for anyone with Internet access to use it. This tool has substantial potential for reaching a significant portion of the population.
In the analysis of the results, if the answer provided by the AI coincided completely with the corresponding content in the protocols, we coded it as ‘right’, and if the answer did not coincide with the protocols or was incomplete, we coded it as ‘wrong’. The level of agreement was calculated as the proportion of right answers in relation to the total number of questions submitted to the AI.
If the response of the AI was “es necesario que acuda a urgencias inmediatamente” (you need to go to the emergency department immediately) or “necesita una valoración médica inmediata” (you need immediate medical assessment), it was interpreted as “need to visit the emergency department”, and if the response was “debe acudir al médico” (you should go to the doctor) or “lleve a su hijo al médico” (take you child to the doctor) it was interpreted as “no need to visit the emergency department”.
We asked the AI about 101 diseases, selected based on their severity and frequency in the months ranging from December 2023 to January 2024. The study was based on previously healthy patients who did not require chronic medication.
Table 1 presents the symptoms of each disease about which we submitted questions to the AI. For each symptom (unless it applied only to a specific age group, in which case it is so noted), we asked about 2 ages: 3 months for infants, and 4 years for young children. In the case of diseases defined in a specific age range (limp in children aged 6-8 years, limp in adolescents) we asked about the specified age.
| Table 1. Symptoms contributed to artificial intelligence | |
|---|---|
| Condition | Symptom(s) |
| Abdominal pain | Tummy ache |
| Acute bronchiolitis | Breathing difficulty |
| Acute dysphagia | Unable to swallow for several hours |
| Acute myositis | Calf pain and inability to walk |
| Acute scrotum | Testicular pain |
| Acute sinusitis | Nasal discharge and headache with or without fever |
| AGE | 4 vomiting and 4 diarrhoea |
| Allergy | Hives after ingestion |
| Allergy | Vomiting after ingestion |
| Allergy | Difficulty breathing after ingestion |
| Anxiety | Nervousness |
| Arthritis, monoarticular | Knee pain and swelling |
| Arthritis, polyarticular | Pain and swelling in hands, wrists and knees |
| Asthma | Respiratory distress |
| Asthma | Cough |
| Asthma | Chest tightness |
| Bite | Dog/cat bite |
| Bradycardia | Slow heart rate |
| Breath-holding spell | Cyanosis and fainting with crying in an infant |
| Bullous rash | Skin blisters |
| Burn | Boiling water burn |
| Burn | Frying pan burns |
| Cardiorespiratory arrest | Loss of consciousness with absence of breathing |
| Cellulite | Erythematous swelling around a wound |
| Chest pain | Chest pain |
| CMPA | Vomiting upon introducing cow’s milk |
| CMPA | Blood in stools upon introducing cow’s milk |
| CMPA | Irritability upon introducing cow’s milk |
| Cold | Cough, nasal discharge and fever |
| Coma/decreased consciousness | Loss of consciousness with impaired breathing |
| Conjunctivitis | Red eye |
| Constipation in infants | 4 days without bowel movements |
| Constipation in preschool-/school-aged children | 4 days without bowel movementsa |
| Cutaneous mycosis | Red patch on one foot |
| Cyanosis | Blue or purple discoloration of lips |
| Dental pain | Pain in a tooth |
| Dental phlegmon | Toothache with swelling of the face |
| Diarrhoea | 6 bowel movements in the past day |
| Diplopia | Seeing double |
| Epistaxis | Nosebleeds |
| Febrile respiratory infection | Cough, mucus and fever up to 39 °C |
| Fever | Fever |
| Foreign body, ear | Sticking a coin/chickpea in the ear |
| Foreign body, gastrointestinal | Swallowing a coin |
| Foreign body, gastrointestinal | Swallowing a battery |
| Foreign body, nose | Sticking a coin/chickpea up the nose |
| Foreign body, ocular | Getting sand in the eye |
| Foreign body, respiratory | Swallowing a coin and difficulty breathing |
| Foreign body, throat | Stuck fish bone |
| GOR in infant | Pours milk at all feedings |
| GOR in older child | Chest pain |
| GOR in older child | Sensation of food coming back up |
| Gross haematuria | Blood in urine |
| Haematemesis | Vomiting blood |
| Headache | Headache |
| Impetigo | Yellow crusts on the skin |
| Inborn error of metabolism | Vomiting in infants |
| Inborn error of metabolism | Seizures in infants |
| Inborn error of metabolism | Decreased level of consciousness or muscle tone in infants |
| Increased intracranial pressure | Severe headache |
| Increased intracranial pressure | Seeing double/paralysis of one side of the face |
| Insect bite | Insect bite |
| Irritability | Inconsolable crying |
| Laryngitis | Hoarse cough |
| Laryngitis | Stridor |
| Leukaemia | Fatigue |
| Leukaemia | Frequent bruising |
| Leukaemia | Difficult to control bleeding |
| Leukocoria | Reflection of flash light is not red |
| Limp in adolescents | Limp |
| Limp in children aged 6-8 years | Limp |
| Limp in toddler/preschooler | Limp |
| Lower gastrointestinal bleeding | Bloody stools |
| Lymphadenopathy | Lump in the neck or groin |
| Mumps | Swelling of the face |
| Neonatal jaundice | Yellowish skin colour in infants |
| Nephrotic syndrome | Eyelid swelling |
| Non-neonatal jaundice | Yellow eyes in children |
| Otalgia | Earache |
| Palpitations | Stabbing chest pain |
| Paronychia | Painful red swelling around nail |
| Pharyngitis/tonsillitis | Fever and sore throat |
| Physical abuse | Bruising on thighs |
| Physical abuse | Decreased level of consciousness in infant (unresponsive) |
| Physical abuse | Cigarette burns |
| Poisoning | Accidental toxic substance ingestion |
| Rash, febrile maculopapular | Red rashes with fever |
| Rash, purpuric | Red spots on the skin |
| Rash, vesicular | Vesicles on the skin |
| Respiratory allergy | Rhinitis |
| Respiratory allergy | Conjunctivitis |
| Scabies | Itching |
| Scabies | Itchy and blotchy skin |
| Seizure, afebrile | Seizure without fever |
| Seizure, febrile | Seizure and fever |
| Seizure, partial | Myoclonus, unilateral arm jerking |
| Sexual abuse | Lesions in vulvar region |
| Sexual abuse | Lesions in anal region |
| Shock | Tachycardia (very fast heartbeat) and pallor |
| SIDS | Loss of consciousness with absence of breathing |
| Stye | Eyelid lump |
| Suicidal ideation | Child says he wants to commit suicide |
| Suppurative AOM | Ear discharge |
| Syncope | Loss of consciousness |
| Tachycardia | Fast heart rate |
| TBI in infant | Fall from changing table |
| Tics | Involuntary movements of the hands or face |
| Torticollis | Neck pain and difficulty moving the neck |
| Torticollis with fever | Pain/difficulty moving neck with fever |
| Trauma, abdominal | Abdominal pain after falling off a bicycle |
| Trauma, ankle | Ankle pain after a fall |
| Trauma, dental | Toothache after a fall |
| Trauma, forearm | Wrist pain after fall |
| Trauma, high-energy | Fall from height/car accident |
| Type 1 diabetes | Frequent urination |
| Type 1 diabetes | High intake with weight loss |
| Urticaria | Skin rashes |
| UTI | Burning sensation in passing urine |
| UTI, febrile in older child | Burning sensation in passing urine |
| Vertigo | Dizziness and spinning of objects |
| Vomiting | Vomiting |
| Wounds | Injury from a fall |
| AGE: acute gastroenteritis; AOM: acute otitis media; CMPA: cow’s milk protein allergy; GOR: gastro-oesophageal reflux; SIDS: sudden infant death syndrome; TBI: traumatic brain injury; UTI: urinary tract infection. | |
For 5 diseases, we asked the AI in regard to all age groups (by month between ages 1 and 24 months and by year for ages 2 to 24 years), and found no differences in the responses of the AI in relation to age. For this reason, we decided to limit questions to age 3 months in reference to infants and age 4 years in reference to young children (early childhood and school age).
We used the Cohen kappa correlation coefficient to measure the level of agreement, which was adjusted for chance in the case of dichotomous variables (need/no need to visit emergency department). For variables that were not dichotomous (recognition of disease, appropriate identification of warning signs and measures to take at home), due to the lack of previous data, we calculated the correlation coefficient without adjusting for chance, using an expected agreement by chance at 0.10 for the calculation of confidence intervals, although the actual probability is likely lower. For diseases for which there are no measures to be taken at home because they always require immediate medical attention, we did not analyse the variables “concerning the assessment of warning signs and the measures to be taken at home. The correlation coefficients and confidence intervals were calculated with the software package SPSS version 28.
We considered the agreement near-perfect if the coefficient was greater than 0.8, substantial if it was between 0.6 and 0.8, moderate if it was between 0.4 and 0.5, fair if it was between 0.2 and 0.4 and poor if it was less than 0.2. If the confidence interval extended across agreement levels, we included both levels in the results.
RESULTS
We analysed responses for 101 diseases, selected based on their frequency and severity. For each disease, we analysed the following variables: recognition of the disease based on the submitted symptoms (Table 1), appropriateness of the warning signs identified by the AI, assessment (correct or incorrect) of the need to visit the emergency department and measures to be taken at home.
The AI assessed the need to seek medical care for each of the symptoms.
Table 2 presents the overall results, Table 3 the results for the warning signs by age group and Table 4 the results by disease category.
| Table 2. Overall results | ||
|---|---|---|
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.857 (0.002) | Near perfect |
| Correct warning signs | 0.888 (0.003) | Near perfect |
| Need to go to emergency room | 0.876 (0.005) | Near perfect |
| Measures to be taken | 0.915 (0.003) | Near perfect |
| Table 3. Age-adjusted analysis, warning signs | ||
|---|---|---|
| Age | Kappa coefficient (95% confidence interval) | Agreement |
| <3 months (infant) | 0.667 (0.053) | Substantial |
| ≥3 months (child) | 0.938 (0.004) | Near perfect |
| Table 4. Results by disease group | ||
|---|---|---|
| Respiratory diseases: asthma, bronchiolitis, upper respiratory tract infection with and without fever, cyanosis | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 1 (0.045) | Near perfect |
| Correct warning signs | 1 (0.053) | Near perfect |
| Need to go to emergency room | 0.733 (0.102) | Substantial |
| Medidas a tomar | 0.778 (0.064) | Substantial |
| Gastrointestinal diseases: neonatal jaundice, non-neonatal jaundice, acute dysphagia, gastro-oesophageal reflux in infant, gastro-oesophageal reflux in older child, haematemesis, lower gastrointestinal bleeding, diarrhoea, vomiting, constipation in infant and constipation in older child. | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.933 (0.021) | Near perfect |
| Correct warning signs | 0.909 (0.024) | Near perfect |
| Need to go to emergency room | 1 (0.04) | Near perfect |
| Measures to be taken | 0.909 (0.026) | Near perfect |
| Dermatological diseases and rashes: maculopapular rash with fever, vesicular rash, bullous rash, purpuric rash, impetigo, cellulitis, paronychia, scabies and cutaneous fungal disease. | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.636 (0.029) | Substantial/Near perfect |
| Correct warning signs | 0.8 (0.032) | Substantial |
| Need to go to emergency room | 0.733 (0.051) | Substantial |
| Measures to be taken | 0.8 (0.032) | Substantial/Near perfect |
| Surgery and trauma: high-energy trauma, abdominal trauma, ankle and foot trauma, forearm trauma, burn, bite, insect bite, wound, gastrointestinal foreign body, respiratory foreign body, torticollis, dental pain | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 1 (0.019) | Near perfect |
| Correct warning signs | 1 (0.026) | Near perfect |
| Need to go to emergency room | 1 (0.031) | Near perfect |
| Measures to be taken | 1 (0.019) | Near perfect |
| Neurologic diseases: irritability, coma, syncope, diplopia, SIDS, increased intracranial pressure, headache, tics, febrile seizure, afebrile seizure, partial seizure, breath-holding spell, acute myositis | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.928 (0.023) | Near perfect |
| Correct warning signs | 1 (0.035) | Near perfect |
| Need to go to emergency room | 1 (0.042) | Near perfect |
| Measures to be taken | 1 (0.032) | Near perfect |
| Oncological diseases: leukocoria, leukaemia | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.25 (0.032) | Weak |
| Need to go to the emergency room | 1 (0.267) | Substantial /Near perfect |
| Allergies: CMPA; food allergy; respiratory allergy | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.25 (0.040) | Weak |
| Correct warning signs | 1 (0.162) | Near perfect |
| Need to go to emergency room | 1 (0.174) | Near perfect |
| Measures to be taken | 0.667 (0.107) | Weak/Substantial |
| Cardiovascular diseases: chest pain, palpitations, bradycardia, tachycardia | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 1 (0.080) | Near perfect |
| Need to go to emergency room | 0.333 (0.129) | Weak/Moderate |
| Measures to be taken | 1 (0.107) | Near perfect |
| Infectious diseases: lymphadenopathy, torticollis with fever, dental phlegmon, acute mumps, fever without source, limp in preschool-aged children, limp in school-aged children, limp in adolescents | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.75 (0.040) | Substantial |
| Correct warning signs | 0.583 (0.026) | Moderate |
| Need to go to emergency room | 0.428 (0.073) | Weak/Moderate |
| Measures to be taken | 0.833 (0.053) | Substantial/Near perfect |
| Ophthalmological and ENT diseases: ocular foreign body, conjunctivitis, stye, foreign body in nose, foreign body in ear, otalgia, suppurative otitis, acute pharyngitis, acute laryngitis, acute sinusitis, foreign body in pharynx, vertigo | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 1 (0.023) | Near perfect |
| Correct warning signs | 1 (0.024) | Near perfect |
| Need to go to emergency room | 0.795 (0.039) | Substantial/Near perfect |
| Measures to be taken | 0.923 (0.023) | Near perfect |
| Metabolic and endocrine diseases: inborn errors of metabolism and type 1 diabetes | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0,8 (0.064) | Substantial/Near perfect |
| Need to go to emergency room | 1 (0.080) | Near perfect |
| Genitourinary diseases: febrile UTI, febrile UTI in older children, nephrotic syndrome, macroscopic haematuria | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.75 (0.080) | Substantial/Near perfect |
| Correct warning signs | 1 (0.080) | Near perfect |
| Need to go to emergency room | 1 (0.129) | Near perfect |
| Measures to be taken | 1 (0.080) | Near perfect |
| Rheumatologic diseases: monoarticular and polyarticular arthritis | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.5 (0.162) | Weak/Substantial |
| Correct warning signs | 1 (0.162) | Near perfect |
| Need to go to emergency room | 1 (0.267) | Substantial/Near perfect |
| Measures to be taken | 1 (0.162) | Near perfect |
| Psychiatric disorders, maltreatment and poisoning: physical abuse, sexual abuse, accidental poisoning, anxiety and suicidal ideation | ||
| Variable | Kappa coefficient (95% confidence interval) | Agreement |
| Recognition of the disease | 0.7 (0,035) | Substantial |
| Correct warning signs | 1 (0,045) | Near perfect |
| Need to go to emergency room | 0.8788 (0,030) | Near perfect |
| Measures to be taken | 1 (0,040) | Near perfect |
| AGE: acute gastroenteritis; AOM: acute otitis media; CMPA: cow’s milk protein allergy; ENT: ear, nose, throat; GOR: gastro-oesophageal reflux; SIDS: sudden infant death syndrome; TBI: traumatic brain injury; UTI: urinary tract infection | ||
DISCUSSION
Although the level of agreement of the AI for the recognition of diseases was 0.86, it proved more effective in identifying warning signs, with a kappa coefficient of 0.89, and in determining whether or not there was a need to visit the emergency department, with a kappa coefficient of 0.88 overall, in addition to giving advice to parents, with a kappa coefficient of 0.91.
Although the overall agreement was substantial, there were large differences in the stratified analysis. While the level of agreement was near-perfect in establishing the diagnosis and the level of urgency common diseases (respiratory, gastrointestinal, ophthalmological and otorhinolaryngological diseases and trauma), the obtained kappa coefficients were lower, although still good, for less prevalent conditions or those with less specific symptoms (endocrine, metabolic, cardiovascular, oncological and rheumatological diseases, psychiatric disorders and suspected abuse).
We ought to specifically comment on dermatological complaints which, on account of their very nature, are difficult to describe with words. In spite of this, the AI achieved a substantial level of agreement in the assessment of warning signs and of the need to visit the emergency department. Something to be considered is the possibility of submitting photographs to AI to increase its diagnostic yield.
When it came to warning signs, AI exhibited near-perfect agreement in children aged more than 3 months. In infants under 3 months, the agreement in clinical warning signs decreased considerably (from 0.94 to 0.67) due to the nonspecificity of symptoms in infants and the fact that, due to their young age, these patients are at increased risk of complications. Thus, when it comes to infants under 3 months, AI cannot be considered a reliable enough instrument to recommend its widespread application.
We did not find any other articles in the literature analysing the effectiveness of artificial intelligence in detecting warning signs compared to the use of protocols and scientific evidence, we were unable to compare our findings with those of previous studies.
We may conclude that AI is a useful tool for classifying symptoms as urgent versus less urgent in children older than 3 months, but it should not replace medical consultation, as doing so could result in missing diseases that, although not requiring immediate medical care, may be serious and difficult to detect, such as oncological diseases or child abuse.
CONCLUSION
The text-based AI performed with a substantial level of agreement with respect to paediatric manuals and protocols commonly used to identify diseases based on symptoms, and near-perfect agreement with existing paediatric protocols in determining the need to visit an emergency department, assessing warning signs and making therapeutic recommendations. We found the highest levels of overall agreement for respiratory, gastrointestinal, trauma/surgical, genitourinary, neurologic, ophthalmological and otorhinolaryngological conditions. The correlation coefficients for these groups were greater than 0.7 in every analysed category.
On the other hand, we found the lowest levels of agreement in the recognition of oncological, allergic and rheumatological diseases, although the AI was effective in recognising the warning signs requiring a visit to the emergency department for all these conditions. In addition, the agreement of AI with standard guidelines and protocols decreased significantly for patients aged less than 3 months. Further studies are needed to assess the performance of artificial intelligence compared to the judgment of a health care professional.
CONFLICTS OF INTEREST
The authors have no conflicts of interest to declare in relation to the preparation and publication of this article.
AUTHORSHIP
All authors contributed equally to the development of the published manuscript.
ABBREVIATIONS
AEPap: Asociación Española de Pediatría de Atención Primaria · AI: artificial intelligence · SEUP: Sociedad Española de Urgencias Pediátricas.
REFERENCES
- Martín R, Sánchez M, Teruel C. Aspectos psicosociales relacionados con la hiperfrecuentación en las consultas pediátricas de atención primaria. An Pediatr (Barc). 2019;90(1):26-31. https://doi.org/10.1016/j.anpedi.2018.03.008
- Mata D, Fernández M, Cobo MV. Hiperfrecuentación en Pediatría: ¿Se trata de una verdadera necesidad? Bol Pediatr. 2015;55(234):243-6.
- Mancera J, Muñoz F, Paniagua F, Gómez F, Fernández ML, Blanca FJ. Problemas de salud y factores determinantes del número de visitas a demanda en pacientes hiperutilizadores de un centro de salud. Aten Primaria. 2001;27(9):658-62. https://doi.org/10.1016/S0212-6567(01)78876-2
- Liang H, Tsui BY, Ni H, Valentim CCS, Baxter SL, Liu G, et al. Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence. Nat Med. 2019;25(3):433-8. https://doi.org/10.1038/s41591-018-0335-9
- Ramgopal S, Sanchez-Pinto LN, Horvat CM, Carroll MS, Luo Y, Florin TA. Artificial intelligence-based clinical decision support in pediatrics. Pediatr Res. 2023;93(2):334-41. https://doi.org/10.1038/s41390-022-02226-1
- Jiménez Alés R. Inteligencia artificial. Desafíos y preocupaciones. Rev Pediatr Aten Primaria. 2023;25:205-10.
- Sociedad Española de Urgencias de Pediatría (SEUP). Protocolos diagnósticos y terapéuticos en urgencias de pediatría. (4th ed.); 2019 [online] [accessed 29/10/2024]. Available at https://seup.org/wp-content/uploads/2024/04/0_Primeras_4ed.pdf
- Asociación Española de Pediatría de Atención Primaria. Algoritmos AEPap [online] [accessed 29/10/2024]. Available at https://algoritmos.aepap.org/
Comments
This article has no comments yet.