The study is a prospective study with three steps: investigating the value of the home videos for diagnosis on the patients’ first visit, the cost savings for these patients with videos when choosing online consultation, and the type of doctor they should choose.

The trial was conducted following the international rules of good clinical practice. The study was approved by the General Hospital of the People’s Liberation Army. Informed consent was obtained from each patient’s parents. The parents would describe the paroxysmal events to the senior epileptologist, who would edit the description to meet the criteria of the clinical medical record and include any relevant personal or family history. Therefore, each patient’s description information was complete and would accurately simulate the patient’s consultation process in the outpatient clinic.

Twelve video recordings showing various paroxysmal events in infants were collected from the Chinese PLA General Hospital outpatient clinic from May 2015 to January 2016. The inclusion and exclusion criteria are listed in Textbox 1. Similar videos were removed until the shortest remained, but we ensured that each typical episode was presented. The flowchart for choosing the videos is presented in Figure 1. The VEEG reports of all 12 patients were also collected.

All corresponding descriptions, home videos, and VEEG reports were presented to two senior epileptologists blind to the study purpose, and they made diagnoses accordingly. Events were categorized as epileptic or nonepileptic: if epileptic, the specific seizure type was listed; if nonepileptic, a diagnosis to explain the paroxysmal events was given. When the diagnoses from the two epileptologists were not the same, a third epileptologist would review the data and provide the diagnoses. We did not encounter a situation in which all three reviewers could not achieve an agreement.

We conducted our investigation in six centers during the pediatric academic conferences. The person playing the videos was unaware of the diagnosis. Figure 2 shows the three steps of the study. A questionnaire was designed to simulate the process of a clinic consultation. The first part of the questionnaire was the basic information of the doctors, including their educational background, profession (pediatrician or pediatric neurologist), working hospital levels (first/secondary/tertiary hospital), and working years. This was because, in clinical work, the doctor’s basic information is open to all patients visiting the outpatient clinic. The second part provided the doctors with the description of the episodes. This part simulated the process of collecting the medical record at the beginning of the patient’s visit. The description provided all the information that the patient would provide to the clinician when there was not a video. The doctors were required to make their diagnoses and fill in the questionnaire. It should be noted that we originally required the doctor to identify the specific type of epilepsy when they considered it an epileptic seizure, but when we simulated the process in our hospital, we found that the general pediatricians could not clearly identify the types of epilepsy. Therefore, we deleted this part and only required the doctors to identify the episodes as epileptic or nonepileptic. The third part of the questionnaire presented the same descriptions as the second part, but the difference was that in the process of actually collecting these data, the doctors would simultaneously see the corresponding videos before making their diagnoses. Before the investigation, the enrolled doctors were informed of the purpose of the study; all data were anonymous. Only the completed and identifiable questionnaires were eligible for our study.

We calculated the cost savings with the help of home videos on the basis of the contribution videos made on correct diagnoses. In this part, we used the model of online consultation instead of the traditional outpatient visit because this is the increasingly popular choice in China with the popularity of mobile phones [22]. Also, we believed that online consultation is a more reasonable choice considering the difficulties and high costs of getting medical services in China, especially for patients from remote areas [23-25]. The cost savings by online consultation and for patients with or without videos were obtained from published articles or government work reports, including transportation fees, accommodation fees, loss of wages, and registration fees. Figure 3 illustrates this simulation process and how the cost savings were calculated.

All data were recorded in Epidata and exported to SPSS version 24 for statistical analysis. Calculations for the correct diagnosis percentage for each episode were made, and the difference was compared by chi-square test. For the impact of the doctors’ background on diagnosis, we gave a correct diagnosis a value of 1 and an incorrect diagnosis a value of 0, and then calculated each doctor’s score with and without the help of home videos. We then compared the influence of doctors’ profession, educational background, working hospital level, and working years on accurate diagnosis by multiple linear regression analysis. We further analyzed the scores of each physician on the diagnosis of epileptic seizures and nonepileptic seizures with or without videos to analyze the role of the doctor’s background information in both cases. The statistical tests presented are two-sided; a P value less than .05 was considered statically significant.

All patients, if they consult online for their first visit, can save the cost of transportation, accommodation, and loss of wages. Because there has not been any standard for the charge of online consultation in China, we equated the fee to the traditional outpatient registration fee, so it could not be saved. There are also no data on the number of pediatric patients with epilepsy; therefore, we first estimated the annual incidence and the total number of new epilepsy patients through related literature [12,26-32]. Then, we calculated the total number of new patients annually who presented with a paroxysmal event and were diagnosed as a nonepileptic event according to the ratio of the epileptic events to nonepileptic events in all paroxysmal events in children [33,34]. The proportion of infant patients was not available, so when calculating the total cost savings of bringing videos and choosing online consultation, we extended the age to 9 years and calculated the total cost savings for patients younger than 9 years nationwide. In addition, if the patients did not obtain accurate diagnoses on their first visit, they would have to pay additional charges on a later visit. We will never know how many visits these patients needed before accurate diagnosis; therefore, we assumed that they would all achieve an accurate diagnosis on their second visit so the cost savings would be the minimum. We then compared the cost savings of bringing videos with that of just a description to study the value of videos. The formulas are presented in Multimedia Appendix 1.

A total of 452 questionnaires were collected, 301 of which met the criteria (66.6%) and were included in the study. Table 1 shows the demographic characteristics of the whole study sample.

The mean number of correct diagnoses through descriptions only and descriptions with home videos was 7.5 (SD 1.7) of 12 and 8.4 (SD 1.7) of 12, respectively. Details of the percentages of correct diagnoses are shown in Table 2.

For epileptic events, the mean percentage of correct diagnoses with only a description available was 84.9% (SD 5.9%), and with a home video it was 88.8% (SD 3.9%). The difference was statistically significant (P<.001). For nonepileptic events, the mean percentage of correct diagnoses with only a description available was 39.9% (SD 15.9%), and with a home video it was 51.4% (SD 17.5%). The difference was also statistically significant (P<.001).

The means of the transportation fees, accommodation fees, further examination fees, and loss of wages according to the published literature and governmental reports [26-31] were US $290.00 (SD $56.14), US $86.30 (SD $22.54), US $192.90 (SD $61.73), and US $104.70 (SD $90.02), respectively. The annual new pediatric patient (younger than 9 years) population is 197,945 in China. Online consultation could save US $1.28 million (8.22 million yuan) per year. For infants with paroxysmal events with videos on their first visit, at least 12.02% of the total cost can be saved.

In the analyses of doctor’s background information, we found that the number of working years was the key factor for a correct diagnosis. Whether or not videos were provided, the level of the doctor’s hospital and educational background were irrelevant to the correctness of the diagnosis. When only parents’ descriptions were available, profession was the only factor that affected the correctness of the diagnosis for epileptic seizures; for nonepileptic seizures, pediatric neurologists with relatively longer working years made more accurate diagnoses. When both descriptions and videos were available, profession and hospital level were the two factors that affected the correctness of the diagnosis for epileptic seizures; for nonepileptic seizures, working years and profession were related to the accuracy of the diagnosis. The statistical results are listed in Table 3.