To understand the effectiveness of AMCs as a revolutionary way of healthcare service delivery, we have conducted a study with the UFH group.Study methodologyFor this study, 200 patients were randomly selected from UFH’s Out-patient department, half of the patients were using AMC-related telehealth consultations and follow-up processes (AMC group), while the other half were using Usual Care delivery methods (UC group), such as in-person outpatient visit and on-site follow-ups. An online survey was given on our telehealth mobile platform, regarding treatment effects based on waiting time and patient satisfaction, with the AMC group versus the UC group. Specifically, we have developed surveys on patient health treatment satisfaction (PSQ18 Questionnaire13) and Patient Willingness to use AMCs.The study was designed to be a Retrospective Observational design14 from the existing patients from UFH Hospitals and Clinics (Beijing, Shanghai, Qingdao, Guangzhou, Shenzhen). The AMC group process consists of telemedicine consultation, mobile heart rate, body temperature detection, mobile diagnosis, and mobile follow-up. In contrast, the UC group process utilizes in-person hospital visits for treating related diseases.We had an AMC mock room for telemedicine consultation, using a pad-screen will 4 g/5 g or clinic Wi-Fi setting, along with wearable monitoring devices, including heart rate, and body temperature detection. A nurse will also sit in the mock room to assist the patient in finishing the evaluation process (such as self-test on heart rate, and body temperatures using wearable monitoring devices).Study follow-up time has been set to be three months in this study, participants in this study provided the above surveys and objective process outcome three months after treatment deployment. Note that patients younger than 18 years and patients not keen to participate in medical-related surveys were excluded from this study based on the methodology presented in15.On the sample size selection, as we pilot with 200 patients with an expected effect size of 1.5, our minimum study Sample size will be 152 subjects to fulfill an a = 0.05 and 80% statistical power requirements. Reversely, a 200 sample size we tend to include in this study, will perform 90% of Statistical Power from the previous setting on margin of error and confidence interval. The Sample size calculator we used is Microsoft Sample Size Calculator TM16. Sample size calculation should follow the statistic Formular: n = (Zα/2 + Zβ)2 *2*σ2 / d2, such that Zα/2 is the critical value at α/2 normal distribution, which is a 95% confidence interval and 0.05 α, whereas Zβ is the critical value at a β 0.2 level, presenting a power of 80%, also σ2 is the population variance, and d is the expected difference17.Study resultsTable 1 Demographics and chariteristics of subjects.From Table 1, we summarized the demograpics of the study. There are 18% of subjects are located at age before 30, 23% are 31–40 years old, 21% are 41–50, 28% are 51–60, 10% are greater than 60 years old. The mean age of subjects is 44, median is 45. 52% of study subjects are male, 48% are female. There are 26% of patients hold a postgraduate degree, 33% have undergrad, 21% college level, 14% high school level, and 6% primary level. Social Economic Sataus are classified as annully income in USD. There have 20% of patients greater than 200,000 income annualy, 39% between 100,000 and 200,000, 23% between 50,000 and 100,000, 13% between 20,000 and 50,000, only 5% below 50,000.Fig. 3Treatment effect on different waiting times: AMC versus Usual Care.Illustrated in Fig. 3, the X axis is the different times on processing treatment (both the AMC group and the UC group), Y axis is the patient reporting treatment effect (by telehealth mobile app survey). The blue line is the Patient’s expected treatment effect on the AMC group by different processing times, while the orange line is the respective outcome of the UC group.From the results, the patient’s expected treatment effect decreases as the processing time increases. From the blue line, when processing time or waiting time on treatment is greater than 52 min, the patient expected AMC treatment effect will decrease to less than 50%. For the UC group, 50% of the expected treatment effect by patients will only have 30 min of waiting time.There is greater than 20 min tolerance by the patient when using AMC compared with Usual care, on their expectation of having 50% treatment effect. Meanwhile, from the result, we can see that AMC has an all-time higher expected treatment effect by the patient, compared with Usual Care. However, when treatment processing time is greater than 60 min, we see the gap between AMC versus Usual Care narrow. In summary, for various health conditions with different treatment times, AMCs deliver much higher treatment effects compared to Usual Care.The next survey we conducted was the association between patient health treatment Satisfaction (PSQ18 Questionnaire) and Patient Willingness to use AMC. Results have been summarized in Fig. 4.Fig. 4Association on patient health treatment satisfaction (PSQ18 Questionnaire) versus patient willingness to use AMC.As the PSQ-18 Score and Patient Willingness to use AMC are both numerical variables, we carried out a Simple Linear Regression analysis to evaluate the association.From the above figure, we can see the Patient Satisfaction Treatment Process (we use the WHO-validated PSQ18 questionnaire), is positively associated with the Patient Willingness to use AMCs. The results indicate that when the Probability of Patient willing to use AMC increases, the patient satisfaction with the treatment process will increase also. In the same way, when AMC’s willingness decreases, Patient Satisfaction also decreases. On the typical cut-off with an 80% Patient Satisfaction Score, we found out that those subjects will also have a roughly 80% probability of using AMC as their healthcare treatment process. Alternatively, such results reflect that AMC will enhance patient satisfaction, not just by decreasing processing and waiting time on health access.Table 2 Association on patient satisfaction under Multivariate Regression Analysis.We also carry out multivariate regression analysis to find out associations between Patient Satisfaction and AMC willingness usage, as well as other covariates and confounders such as Education level, Social Economic Status, age and gender. Education levels are graded in 5 classes, from Postgrad, Undergrad, College, High school to Primary school Level. Social Economic Status are classified as annually income, in USD.From result of Table 2, we found that AMC willingness usage is associated with patient satisfaction in a positive direction, on (coefficient 77.78, P value < 0.01). Showing that patient keener to use Autonomous Mobile Clinic consultation would have a higher score on patient satisfaction. While age is associated with patient satisfaction in a negative direction. (coefficient − 0.41, P value 0.01). Education level and Social Economic Status shown positive association with patient satisfaction, on coefficient 8.83, P value < 0.01, and coefficient 6.96, P value < 0.01, respectively. Results for all factors are adjusted for other covariates. (e.g. AMC willingness result is adjusting for Age, Gender, Education level, Social Economic Status. Other results of factors are adjusted, respectively).
