Venous thrombosis, which includes deep vein thrombosis, pulmonary embolism, and cardiogenic stroke, poses a significant burden on the health care system [1]. Anticoagulants are effective in the treatment of venous thrombosis. Patients with atrial fibrillation and valve replacement surgery are at high risk for venous thrombosis; therefore, these patients require anticoagulant treatment. Warfarin is the most commonly used oral anticoagulant due to its low price, specific antagonists, and substantial evidence-based clinical data [2]. The dose of warfarin required for anticoagulation is closely related to dietary vitamin K intake, concurrent medications, the presence of other diseases, body weight, and aerobic exercise. Therefore, warfarin doses during anticoagulation therapy need to be adjusted according to the results of international normalized ratio (INR) tests. Without proper adjustment, excessive warfarin may cause bleeding, and inadequate warfarin doses may cause thrombosis and possible death [3]. This represents a costly use of medical resources and is physically and psychologically traumatic to patients and their families [4].

Warfarin dose adjustment is usually performed at hospital clinics. For patients living in rural areas, adjustment of warfarin dosages requires transportation, accommodation, and time. Patients with low incomes are willing to risk thrombosis and bleeding in an attempt to reduce costs. Online services such as smartphones, text messages, Bluetooth, and communication platforms may assist in adjusting warfarin doses [4-6]. Our meta-analysis, which included 16,915 patients, of whom 8655 had their warfarin dosage adjusted in the hospital and 8260 had their warfarin dosages adjusted through online services, revealed that online services were associated with fewer warfarin-related hospital admissions than hospital management (odds ratio 0.47, 95% CI 0.30-0.73; P<.001). However, there was no statistically significant difference in other anticoagulant effectiveness or clinical outcomes between the two groups [4]. Our previous research provided online services for patients who were taking warfarin through QQ group communication platforms, which can be accessed via smartphones, computers, laptops, and tablets. QQ is an instant messaging software service developed by Chinese company Tencent Holdings Limited. As described by You et al, social groups are some of “the main features of QQ [that allow] multiple users to communicate instantly. A message posted by a member is immediately received by all the other group members” [7]. The results showed that online services yielded similar clinical outcomes to hospital services, even though the incidence of supratherapeutic INR values increased [6]. However, there were some limitations in the QQ group communication platform. First, one patient reported his medical condition, and all the other patients could see it. There was no method of protecting patient privacy. Second, it was challenging to obtain the most recent medical information of patients because the information on all patients was mixed. Third, the information could only exist temporarily.

Recently, we developed a new warfarin dosage adjustment app named Alfalfa, which is available for installation on a mobile device; the communication between the physicians or pharmacists and the patient is point-to-point. The privacy of the patients is well protected. Meanwhile, personal medical information can be retrieved by the patients or their physicians or pharmacists at any time. As far as we know, apart from Alfalfa, there are only two studies focusing on managing warfarin with the use of mobile apps that can automatically suggest a dose and time for the next blood test based on the patient's INR value [8,9]. The difference between these mobile phone apps and Alfalfa is that Alfalfa provides a channel for communication between patients and medical staff, while the other two warfarin management apps focus on patient self-management. Alfalfa consists of a background management system and a remote anticoagulation management system. This remote anticoagulation management system can be divided into a patient terminal called Alfalfa Health Management and a medical terminal called Alfalfa Anticoagulation Guidance. The main function of Alfalfa is to facilitate the exchange of information between patients and medical staff, so that patients can report blood coagulation results and changes in their health status to physicians or pharmacists through the internet and then take their medicine and check the INR value of warfarin according to the advice of physicians or pharmacists [10]. In this study, we explored the effectiveness and safety of warfarin dose management via the Alfalfa app, which can provide further evidence in support of anticoagulant management via online anticoagulation services.

Patients taking warfarin need to monitor INR and adjust their warfarin dosages accordingly. The included patients took warfarin mainly for the following three reasons: heart valve replacement, vein thromboembolism, and atrial fibrillation. Warfarin is the only oral anticoagulant available for thrombosis prevention after heart valve replacement [15]. The type of prosthetic valve, its anatomical location, and patient-specific risks of thromboembolism and bleeding influence the specific intensity and duration of antithrombotic treatment to prevent prosthetic valve thrombosis [16]. Some studies have proved that Chinese people need lower anticoagulation intensity INR (1.5-2.5) to warfarin in comparison to the recommended INR (2.5-3.5) in developed countries [17-19]. As long-term anticoagulant therapy for vein thromboembolism, an authoritative guideline has suggested warfarin adjusted to achieve an INR of 2.0-3.0 [20]. In patients with atrial fibrillation, warfarin reduces the relative risk of stroke by 64% and all-cause mortality by 26% [21]. Clinical studies have confirmed that when INR ranges between 2.0 and 3.0 in patients with atrial fibrillation, warfarin effectively prevents stroke and does not significantly increase the risk of bleeding [22]. The effectiveness of anticoagulant therapy is usually expressed as TTR or INR within the therapeutic target range. TTR values greater than 65% are indicative of effective anticoagulation therapy [21]. Due to the different medical standards and management methods in different regions, the TTR range is wide (29%-75%) [23]. Even in strictly controlled, large-scale clinical trials such as the ROCKET AF study, the average TTR was only 55.2% for all patients, among whom Chinese patients with atrial fibrillation had a TTR of only 47%. It was challenging for anticoagulation therapy to achieve the desired effect [23,24].

In this study, TTR was significantly higher in the Alfalfa app group than in the offline warfarin management group (79.35% vs 52.38%, P<0.001). These results were similar to those reported by Prochaska et al, who showed that TTR was significantly higher from telemedicine-based coagulation service than from regular anticoagulation management (75.5% vs 66.3%, P<0.001) [25]. However, a study by Lee et al revealed opposite results [26]. Some studies have reported similar TTR between online anticoagulation management and regular medical care [27-29]. This study showed that the rates of subtherapeutic, supratherapeutic, and extreme subtherapeutic INR values were significantly lower in the Alfalfa app group than in the offline warfarin management group. The lower rate of abnormal INR distribution may reduce the incidence of bleeding and thrombotic events.

Our research findings showed that the Alfalfa app had lower incidences of major bleeding events, warfarin-related emergency department visits, and warfarin-related hospital admissions than the offline warfarin management group. However, the incidence of minor bleeding events was higher in the Alfalfa app group. This finding was similar to the results reported by Blissit et al, who concluded that the incidence of minor bleeding events was higher in the online anticoagulation management group [30]. The top three minor bleeding types in the offline group were nose bleeding (7 cases), gum bleeding (7 cases), and hematuria (4 cases). The top three minor bleeding types in the Alfalfa app were gum bleeding (10 cases), menstrual abnormalities (increased, prolonged, or advanced; 7 cases), and subcutaneous purpura (9 cases). More minor bleeding events in the Alfalfa app may be due to the way patients provided feedback. The data of the offline group are obtained through follow-up, while the data of the Alfalfa app group are actively reported by the patient in the app. This study is a retrospective study. Most patients received follow-up calls after taking warfarin for 3 months. It is possible that patients could not recall the minor bleeding adverse reactions and relay them to the investigator. On the other hand, patients in the Alfalfa app group have a stronger willingness to communicate with medical staff. Thus, the Alfalfa app showed more minor bleeding events.

The studies by Xia et al and Cryder et al showed that there were fewer anticoagulation-related hospital admissions in the online anticoagulation group [4,27]. Most studies have reported that there are no differences in the incidence of major bleeding events between the two groups [5,27-32]. Our findings revealed that the incidences of thromboembolic events were similar between the two groups. While the offline anticoagulant management models were similar among the studies, the online management models were slightly different. For example, some studies managed patients through text messages, while others relied on Bluetooth and online communication platforms. The Alfalfa app is the first app specifically designed for warfarin dose adjustment management. This study showed that the application of the app not only achieved better anticoagulant control effects, but also greatly reduced major bleeding events, emergency department visits, and hospitalization. Therefore, the Alfalfa app is likely to reduce related medical costs. There are two main reasons for the positive effects of the Alfalfa app. First, the Alfalfa app can retrieve the patient’s previous warfarin dosage, INR value, bleeding and thrombotic events, and adverse reaction history so that physicians and pharmacists can better analyze the reason for abnormal INR or bleeding. Physicians and pharmacists can assess whether the abnormal INR or bleeding events are caused by warfarin or induced by other diseases. Second, the Alfalfa app has an automatic response function for the physicians or pharmacists according to the protocols of warfarin dose adjustment. After the automatic response, physicians or pharmacists can check whether the automatic response is correct before it is submitted to the patients. This function is considerably helpful for anticoagulation management newcomers.

However, the results of this study showed that anticoagulation control may not be completely consistent with the clinical outcomes. For example, the extreme supratherapeutic INR values in this study were distributed close to each other, but the incidence of major bleeding events was lower in the Alfalfa app group. The average TTR of the offline warfarin management group was only 52.38% (<60%). In theory, the incidence of thrombosis should be higher in the offline warfarin management group, but the incidences of thrombosis were similar between the two groups.

In recent years, the internet has made considerable progress in the management of chronic diseases, especially hypertension and diabetes. The internet provides novel opportunities in anticoagulation management. Telemedicine breaks geographical and spatial limitations, providing high-quality medical resources for patients living in vast rural and remote areas. Similarly, INR real-time detection technology is convenient for at-home INR detection. This technology allows patients to enjoy high-quality online anticoagulation management services at home.

Currently, rural and medically underresourced areas often lack specialist anticoagulant physicians or pharmacists, resulting in a significant risk of bleeding and thrombosis for patients living in these areas. We are conducting a national randomized controlled multicenter study comparing the effects of the Alfalfa app and offline services [33]. If the Alfalfa app becomes widely available, there could be significant benefits for such patients. At the same time, we also look forward to the economic evaluation results of these two management models.

There were some limitations to this study. First, as a retrospective study, subjects were not assigned to either group; therefore, selection bias may be present. Second, the participants were included from a single center in Southeast China, which may not be representative of all patients with anticoagulation therapy. More prospective, randomized, and multicenter studies are required to confirm our findings.

In conclusion, warfarin management in the Alfalfa app group was superior to that in the offline group in terms of TTR, abnormal INR, major bleeding events, warfarin-related emergency department visits, and warfarin-related hospital admissions. Warfarin management via the Alfalfa app may be suitable for patients living in rural and remote areas.