Evaluation of a pharmacist-led patient-self-testing model for warfarin management in patients undergoing mechanical heart valve replacement in China: a multicentre, open-label, randomised, controlled trial

Abstract

Objectives

Patient self-testing (PST) for warfarin management is well-established in developed countries but remains underused in developing regions. This study compared the safety and effectiveness of PST with usual care (UC) in China.

Design

A multicentre, open-label, randomised, controlled trial.

Setting

A total of five centres participated in this study, including one provincial tertiary hospital, two municipal tertiary hospitals and two primary hospitals.

Participants

Patients undergoing mechanical heart valve (MHV) replacement at five centres were prospectively enrolled. Patients were trained and stratified according to time on warfarin at enrolment and were randomly assigned to the PST or UC group.

Interventions

The PST group used a point-of-care testing device for at-home international normalised ratio (INR) monitoring with pharmacist-guided warfarin dosing, while the UC group attended outpatient clinics for INR monitoring and dosing.

Primary and secondary outcome measures

The primary outcome was the difference in time in therapeutic range (TTR). The secondary outcomes were incidences of major bleeding, thromboembolism and all-cause deaths in 12 months.

Results

From March 2021 to March 2023, a total of 556 patients were enrolled, with a mean age of 47.5 years, 45.1% being male. 342 were newly initiating warfarin therapy, while 214 had been on warfarin for over 6 months. Baseline characteristics were similar between the PST and UC groups. The PST group showed significantly higher TTR (67.2% vs 55.1%, p<0.001) and lower incidences of major bleeding (0.7% vs 7.9%, p<0.001) and thromboembolism (0.4% vs 6.8%, p<0.001), with no difference in all-cause mortality (0.4% vs 1.8%, p=0.22). Logistic regression identified that using PST and younger age were independent factors associated with fewer warfarin-related adverse events.

Conclusions

A pharmacist-led PST intervention with ongoing education and counselling led to improved TTR and clinical outcomes in patients with MHV in China.

Trial registration number

China Clinical Trial Registry (ChiCTR2000038984).

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A key methodological strength was the implementation of structured pharmacist-led education and counselling to support patient self-testing (PST).

• The study was designed as a multicentre, open-label, randomised, controlled study comparing PST against usual care.

• The study used a logistic regression analysis to identify factors associated with adverse events.

• The recruitment of participants was limited to hospitals within a single province in China and the time in therapeutic range for the control group was low. This may affect the generalisability of the findings.

• The study eligibility criteria excluded paediatric populations, limiting the applicability of the findings to adults.

Introduction

Millions of individuals worldwide rely on warfarin daily for the prevention and treatment of thrombosis.^{1 2} Despite the widespread adoption of direct oral anticoagulants, warfarin remains the preferred anticoagulant for short-term use in patients with biological heart valves³ and long-term use in those with mechanical heart valves (MHVs).¹ Due to warfarin’s narrow therapeutic window and significant inter-individual variability, it is essential to regularly adjust dosage based on the prothrombin time (PT) and international normalised ratio (INR) values to ensure safety and effectiveness.⁴ In practice, clinicians often use the average time in therapeutic range (TTR) to assess patients’ quality of warfarin therapy.^{5 6}

In China, the predominant model for warfarin management is usual care (UC), where physicians manage a large number of patients, primarily focusing on warfarin dose adjustments. A less common alternative is the anticoagulation clinic (AC), where trained pharmacists adjust warfarin dosages, educate patients and conduct follow-ups. However, ACs are mostly limited to tertiary hospitals. Warfarin management in China is still unsatisfying, where TTR was only 55%, significantly lower than that of the 60%–77% in developed countries.⁷ Clinical data further highlight elevated rates of warfarin-associated ischaemic stroke (5.23%) and intracranial haemorrhage (2.94%) in China, compared with global averages of 2.07% and 0.63%, respectively.⁸ These statistics underscore an urgent need for innovative anticoagulation management strategies in low- and middle-income countries.

In developed countries, the patient self-testing (PST) and patient self-management (PSM) models are widely used.⁹ For PST, patients use a point-of-care testing (POCT) device to measure INR values at home, while physicians or pharmacists remain responsible for adjusting warfarin dosage. In contrast, PSM involves patients taking full responsibility for both INR monitoring and dosage adjustments.¹⁰ Meta-analyses have reported that PST and self-management for warfarin improved TTR, but only PSM improved a clinical outcome (thromboembolic events).¹¹ Meanwhile, a low-quality retrospective study also showed that PST was associated with lower rates of thromboembolism, haemorrhage and emergency department visits compared with office/laboratory-based INR monitoring.⁹ However, its widespread adoption, even in developed countries, is hindered by barriers such as high initial costs and the requirement for robust patient training and support systems.^{12 13}

In China, per capita healthcare resources remain relatively limited. Studies show that China has a lower physician-to-population ratio compared with other nations.¹⁴ Consequently, Chinese physicians, primarily engaged in inpatient and outpatient services, may find it challenging to manage patients on warfarin via the PST model. In contrast, trained pharmacists have demonstrated a positive role in chronic disease management.15,17 So far, no multicentre studies have reported the safety and efficacy of PST in China or other low- and middle-income countries. To address this gap, our team carried out a multicentre randomised controlled trial (RCT) to investigate the safety and effectiveness of the pharmacist-led PST in patients undergoing MHV replacement in China.

Methods

Study design

The detailed protocol was previously reported.¹⁸ This multi-centre prospective RCT was initiated by the Second Xiangya Hospital of Central South University in China. This was an open-label study due to the nature of the warfarin management intervention. We finally enrolled patients from five centres, including Second Xiangya Hospital, Xiangtan Central Hospital, Taoyuan People’s Hospital, the First People’s Hospital of Pingjiang and People’s Hospital of Ningxiang City.

The study protocol was registered with the China Clinical Trial Registry (https://www.chictr.org.cn/). This study followed the Consolidated Standards of Reporting Trials reporting guidelines.

Participants

Eligible patients were those who were 18 years or older with MHV replacement and who signed informed consent. Exclusion criteria were as follows: pregnancy or intention to become pregnant within the next 18 months; history of bleeding complications requiring treatment within the past 6 months; inability to reach a hospital within 6 hours; renal failure (haemodialysis or glomerular filtration rate <10 mL/min); liver dysfunction (Child-Pugh score B or C); dementia or unstable psychiatric conditions; or ongoing cancer treatment.

Additionally, enrolled patients were required to use the POCT device to perform at least two reproducible INR measurements under the supervision of the research staff. These results were compared against laboratory INR measurements from venous blood samples. Under correct operation, patients were excluded from the study if the difference between POCT and laboratory INR values exceeded 0.3 units in order to ensure POCT results’ accuracy.¹⁹

Randomisation

A study suggested that INR typically stabilised around 129 days after initiating warfarin therapy,²⁰ and based on our experience, most patients achieve stable INR after 6 months. Accordingly, we recruited patients from both outpatient and inpatient departments and divided them into two subgroups. Group 1 consisted of patients who had just undergone MHV surgery and had not yet been discharged from the hospital, while group 2 included patients who had completed MHV surgery and had been taking warfarin for more than 6 months. Participants were randomly allocated 1:1 to PST or control groups using sequentially numbered, opaque, sealed envelopes (SNOSE). The randomisation sequence was generated via an online tool (http://stattrek.com/random-number-generator) and concealed in SNOSE. Envelopes were opened after obtaining written consent and recording baseline data. Group assignment was immediately disclosed to patients and pharmacists, while outcome assessors remained blinded throughout the study.

Pharmacists’ training

All clinical pharmacists in this study completed standardised training accredited by the Chinese Hospital Association or the Chinese Medical Association before study initiation. The principal investigator also provided unified anticoagulation training courses to all the participating clinical pharmacists before patients’ enrolment. The training included anticoagulation knowledge and skills (eg, patient assessment, warfarin dose adjustment), use of the POCT device and its accompanying mobile application, patient communication, adherence support and management of adverse events (eg, bleeding or thromboembolism).¹⁸ Periodic refresher sessions were also conducted throughout the study to maintain care quality.

Patients education

All eligible patients received standard warfarin education provided by pharmacists at enrolment, including a face-to-face instructional session, a video and an information brochure.¹⁸ The education covered basic knowledge of warfarin, blood tests of INR and target ranges, follow-up frequencies, recognition of signs and symptoms of over-coagulation and under-anticoagulation, adverse events treatment, drug/food interactions with warfarin, etc. For further details, refer to online supplemental file 1.

Intervention

PST group

After randomisation, patients in the PST group additionally received training of using a POCT device (qLabs ElectroMeter Plus coagulonometer, Micropoint Biotechnologies Incorporated, Guangdong, China)²¹ to monitor PT/INR and operating the accompanying mobile application, including an overview of INR self-testing and the POCT device, proper finger-prick sampling techniques, coagulometer operation and recording and uploading test results to the mobile application on their smartphones. On completing the training, pharmacists assessed patients’ self-testing competency, ensuring proper operation, data recording and result submission. Patients who failed the assessment underwent retraining until proficiency was achieved.

During PST implementation, patients measured their INR values at home and uploaded the results along with their recent health status to the mobile application. On the same day, a trained pharmacist adjusted warfarin dosage as needed and informed the next INR monitoring date. Patients were required to monitor INR at home at least once every 2–4 weeks and were also required to visit a physician and monitor INR through a lab test for an external quality assessment²² of the POCT device every 3–6 months. Based on their INR values and whether presenting adverse events, patients were categorised as ‘safe’, ‘vigilant’ and ‘dangerous’ for different management strategies.¹⁸ Doctors were responsible for treating adverse events, while pharmacists handled patient education and dosage adjustment. Each patient in the PST group was provided with a POCT device, free pharmacist guidance and subsidised testing strips. Pharmacists were also responsible for promptly responding to patient medication-related inquiries via the mobile application or WeChat.

UC group

After randomisation, patients in the UC group visited an outpatient clinic, where venous blood samples were drawn for INR monitoring, and doctors adjusted their warfarin dosage and informed the next INR monitoring date. No medication adjustments or additional patient education were provided by pharmacists to this patient group.

Outcomes

The primary outcome was the difference in TTR during 12 months between the PST and UC groups, which was calculated by Rosendaal’s method,²³ which uses linear interpolation to assign an INR value to each day between two consecutive measured INRs. All INR values obtained during the study period were included in the calculation. An independent endpoint adjudication committee, blinded to group allocation, verified the integrity of all INR data before analysis. The Chinese expert consensus on postoperative anticoagulation following heart valve surgery determines the INR target range as 1.8–2.5 for aortic and mitral valve replacement, and 2.0–3.0 for tricuspid valve replacement or MHV combined with atrial fibrillation, previous cerebral infarction, hypercoagulable state and other special conditions.²⁴

The secondary outcomes included major bleeding and thromboembolic events, all-cause mortality and extreme INR values. Major bleeding was defined as bleeding classified as BARC 2 or higher.²⁵ Thromboembolic events included valvular thrombosis, ischaemic stroke, acute myocardial infarction or arterial thromboembolism.²⁶ For each reported event, relevant medical records were reviewed to assess the clinical characteristics of the bleeding or thromboembolic episode. All dates of death were confirmed by medical records. Extreme INR values were defined as INR<1.5 or INR>5.0.¹⁸

Data collection and follow-ups

At the outset of the study, baseline information was collected from each patient as shown in table 1 and online supplemental table 1. All the patients were followed up at the 3rd, 6th, 9th and 12th months through a combination of outpatient clinic visits, telephone or internet interviews. At each follow-up time point, the researchers collected the monitoring booklets for review, including data on INR results, warfarin dosages, concurrent medications, dietary changes, new illnesses and any thrombotic or haemorrhagic events.

Table 1. Demographics and baseline characteristics of all patients.

Characteristic	Total patients (n=556)	Total PST group (n=278)	Total UC group (n=278)	P value
Age (years)				0.70
Mean±SD	47.5±10.7	47.7±11.8	47.3±9.6
Gender, no. (%)				0.93
Male	251 (45.1)	126 (45.3)	127 (45.7)
Female	305 (54.9)	152 (54.7)	151 (54.3)
Target INR range, no. (%)				0.48
1.8–2.5	350 (62.9)	171 (61.5)	179 (64.4)
2.0–3.0	206 (37.1)	107 (38.5)	99 (35.6)
Years of using warfarin, mean±SD	7.57±5.51	NA	NA
History of adverse events, no. (%)	48 (8.6)	24 (8.6)	24 (8.6)	1.0
Thromboembolism	39 (7.0)	20 (7.2)	19 (6.8)	0.87
Major bleeding	11 (2.0)	5 (4.7)	6 (2.2)	0.76
Comorbidities, no. (%)
Hypertension	65 (11.7)	38 (13.7)	27 (9.7)	0.15
Diabetes	18 (3.2)	10 (3.6)	8 (2.9)	0.63
Coronary heart disease	35 (6.3)	18 (6.5)	17 (6.1)	0.93
Atrial fibrillation	169 (30.4)	86 (30.9)	83 (29.9)	0.78
Other diseases	67 (12.1)	35 (12.6)	32 (11.5)	0.70
Smokers, no. (%)	78 (14.0)	37 (13.3)	41 (14.7)	0.63
Average daily dose of warfarin (mg), mean±SD	3.2±1.0	3.2±1.1	3.2±0.9	0.99

INR, international normalised ratio; NA, not applicable; PST, patient self-testing; UC, usual care.

Sample size calculation

The sample size calculation determined that 31 patients per group would provide 80% power (α=0.05) to identify a 5% TTR difference between the intervention and control groups. Considering a dropout rate of approximately 10%, 34 patients were planned to be enrolled in each group.¹⁸ For the secondary outcome of major bleeding, the anticipated difference between groups may be small. Assuming a 2.94% major bleeding rate in the control group⁸ and 1.84% in the intervention group, the calculated sample size was 233 per group. Allowing for a 10% dropout rate, at least 256 participants were required in each group.

Statistical analysis

All analyses were performed in the intention-to-treat population (all the patients who underwent randomisation) according to trial group. Continuous variables were assessed for normality using the Kolmogorov-Smirnov test before analysis. Normally distributed data were analysed with the independent samples t-test, while non-normally distributed data were analysed using the Mann-Whitney U test. Categorical variables were presented as counts (percentages) and analysed using Pearson’s χ² test or Fisher’s exact test, as appropriate. To evaluate the risk of adverse events, survival curves were generated by the group using the Kaplan-Meier method, with comparisons conducted using the log-rank test. Binary multivariate logistic regression was used to identify risk factors associated with adverse events. A two-sided p value of less than 0.05 was considered statistically significant. All data analyses were performed using SPSS (V.26.0).

Results

Characteristics of the patients

From March 2021 to March 2023, a total of 963 patients were assessed for eligibility, with 556 ultimately meeting the inclusion criteria and undergoing randomisation. Participants were allocated to either the PST or the UC group (n=278 in each group) as depicted in figure 1. In the PST group, seven patients were withdrawn by the investigator due to not being compliant with the protocol. In the UC group, 12 patients withdrew consent, and 11 patients were lost to follow-up. Finally, 271 patients remained in the PST group and 255 patients in the UC group.

Figure 1. Study flowchart. PST, patient self-testing; SNOSE, sequentially numbered, opaque, sealed envelopes; UC, usual care.

The mean age of the total population was 47.5±10.7 years, and 251 (45.1%) were male. The average daily warfarin dose was 3.2±1.0 mg. The study cohorts demonstrated balanced baseline characteristics between the two groups, as shown in table 1. No significant differences were observed between the paired groups in baseline characteristics of the total population and within the subgroups, including age, sex, valve position, comorbidities, smoking status and average daily warfarin dose. For patients in group 2, the mean duration of warfarin use was 7.5±5.5 years in the PST group and 7.6±5.6 years in the UC group. There were no significant differences in the history of adverse events between the PST group 2 and the UC group 2.

Primary outcome

Figure 2A illustrates TTR of the PST and UC groups. Median TTR was significantly higher in the total PST group compared with the total UC group (67.2% vs 55.1%, p<0.001). compared with their respective UC subgroups, median TTR was also significantly higher in PST group 1 (67.2% vs 54.3%, p<0.001) and PST group 2 (67.0% vs 56.0%, p<0.001), as illustrated in figure 2B,C.

Figure 2. Primary outcome and secondary outcome. The violin contour plot demonstrates the time percentage in therapeutic range (TTR). The above graphs showed the total group (A), group 1 (patients newly initiating warfarin, B) and group 2 (patients taking warfarin over 6 months, C). (D–F) Display bar charts illustrating the incidence of adverse events in the total group (D), group 1 (patients newly initiating warfarin, E) and group 2 (patients on warfarin for over 6 months, F). PST, patient self-testing; UC, usual care.

Tables2 3 present the monitoring characteristics of the study. During the follow-up period, the PST group recorded 11 352 INR values, with a median of 41 tests per person and a median testing frequency of every 7 days. In contrast, the UC group recorded 2397 INR values, with a median of 8 tests per person and a median testing frequency of every 29 days (p<0.001). Table 2 also presents the number of extreme INRs reported in both groups, with significantly fewer subtherapeutic and supratherapeutic INRs observed in the PST management group. The PST group had a significantly higher number of tests and a more frequent testing schedule compared with the UC group, which were the main reasons contributing to the increased TTR and reduced incidence of extreme INR values. Subgroup analyses of extreme INR values comparison between PST group 1 and UC group 1, and PST group 2 and UC group 2 were similar as shown in table 3.

Table 2. INR monitoring characteristics of all patients.

	Total PST group (n=278)	Total UC group (n=278)	P value
Number of INR tests per patient			<0.001
Median, (IQR)	41 (29–51)	8 (6–10)
Frequency of testing (days)			<0.001
Median, (IQR)	7 (4–11)	29 (13–68)
Critical INR values, no. (%)
INR>5.0	0.1 (9/11352)	0.7 (16/2397)	<0.001
INR<1.5	6.3 (712/11352)	12.0 (288/2397)	<0.001

INR, international normalised ratio; PST, patient self-testing; UC, usual care.

Table 3. INR monitoring characteristics of patients across different subgroups.

	PST group 1 (n=171)	UC group 1 (n=171)	P value	PST group 2 (n=107)	UC group 2 (n=107)	P value
Number of INR tests per patient			<0.001			<0.001
Median, (IQR)	46 (36–55)	8 (6–11)		32 (22–44)	7 (5–10)
Frequency of testing (days)			<0.001			<0.001
Median, (IQR)	7 (4–10)	28 (12–62)		7 (5–14)	39 (14–83.75)
Critical INR values, no. (%)
INR>5.0	0.1 (1/7727)	0.6 (9/1587)	<0.001	0.2 (8/3625)	0.9 (7/810)	0.01
INR<1.5	6.1 (469/7727)	13.7 (218/1587)	<0.001	6.7 (243/3625)	8.6 (70/810)	0.05

Group 1, patients who had just undergone mechanical heart valve (MHV) surgery and had not yet been discharged from the hospital; group 2, patients who had completed MHV surgery and had been taking warfarin for more than 6 months.

INR, international normalised ratio; PST, patient self-testing; UC, usual care.

Secondary outcomes

In total, the PST group experienced one death (0.4%), two cases of major bleeding (0.7%) and one thromboembolism (0.4%). In contrast, the UC group had 5 deaths (1.8%), 22 cases of major bleeding (7.9%) and 19 thromboembolisms (6.8%). Online supplemental table 2 presents detailed adverse events occurring in each group. Figure 2D–F shows 12-month overall adverse event rates (proportions) in the whole population and subgroups. The incidence of total adverse events was significantly lower in the PST group compared with the UC group (1.4% vs 16.5%, p<0.001). The incidence of all-cause death was similar in these two groups (p=0.22). Compared with the UC group, the incidences of major bleeding (7.9% vs 0.7%, p<0.001) and thromboembolism (6.8% vs 0.4%, p<0.001) in the PST group were both significantly lower.

Figure 3 displays time-to-event dynamics via Kaplan-Meier curves between the two groups over 12 months, accounting for event timing and censoring. The risk of major bleeding in the PST group was significantly lower compared with the UC group (HR=0.17; 95% CI 0.08 to 0.38; log-rank p<0.001), and a similar phenomenon was observed in the thrombotic complications (HR=0.15; 95% CI 0.06 to 0.37; log-rank p<0.001). There was no significant difference in survival rates for all-cause death between the two groups (log-rank p=0.0885).

Figure 3. Kaplan-Meier survival curves. (A–D) Present Kaplan-Meier estimates of event-free survival for the PST and UC models, showing rates for any adverse events (A), major bleeding (B), thromboembolism (C) and all-cause mortality (D). PST, patient self-testing; UC, usual care.

Risk factors related to adverse events

We carried out a univariate analysis (online supplemental table 3) and binary logistic regression analysis (table 4), revealing factors related to adverse events by using the total population. The results indicated that receiving PST exhibited a significantly lower risk of experiencing adverse events (OR=0.067, 95% CI 0.023 to 0.190). Moreover, older age (OR=1.049, 95% CI 1.015 to 1.085) was also an independent factor related to having adverse events.

Table 4. Binary logistic regression analysis revealing factors related to adverse events.

Variables	β	SE	OR	95% CI	P value
Group					<0.001
UC Group	reference
PST group	−2.709	0.535	0.067	0.023 to 0.190
Age	0.048	0.017	1.049	1.015 to 1.085	0.004

PST, patient self-testing; UC, usual care.

Discussion

This study was the first multicentre RCT in China to evaluate the difference between pharmacist-led PST and UC in patients after MHV. We have two main findings. First, compared with patients on warfarin with infrequent follow-ups in the UC group, pharmacist-led PST with ongoing education and counselling led to improved quality of warfarin management, as evidenced by a higher TTR as well as significantly fewer bleeding and thromboembolic events. Second, within this study population, younger patients and those assigned to PST were associated with a lower risk of adverse events.

In the PST group, we observed lower incidences of major bleeding and thromboembolism compared with the UC group, while the rate of all-cause mortality remained similar between the two groups. This finding aligned with the results reported by Beyth et al.²⁷ The differences in INR monitoring frequency between PST and UC groups (41 vs 8 tests/year) accounted for the main reason for the observed outcome disparity. The PST group’s increased testing frequency improved TTR while reducing extreme INR values, collectively driving better clinical outcomes and fewer adverse events.^{28 29} Previous clinical studies have shown that PST can improve TTR,1927 30,32 which was consistent with the results of our study. Additionally, timely warfarin dose adjustments could reduce adverse events in patients,^{33 34} and in this study, we showed that the PST model effectively achieved this goal. Subgroup analyses demonstrated that PST exhibited favourable safety and efficacy profiles in both warfarin new starters and those with prior warfarin experience compared with the UC subgroup patients.

This study demonstrated that in a Chinese clinical setting with suboptimal routine monitoring, pharmacist-led PST was associated with improved anticoagulation control compared with UC, suggesting its potential as an alternative model for eligible patients. According to WHO data,¹⁴ the density of physicians in China in 2021 was 25 per 10 000 population, which was lower than that in developed countries, such as the USA (36 per 10 000) and the UK (32 per 10 000). Furthermore, medical resources in China are more concentrated in economically developed regions.³⁵ Due to high workloads, Chinese doctors prioritise diagnosis and treatment, often with less emphasis on patient education. Currently, Chinese pharmacists have achieved favourable outcomes in managing chronic illnesses, including chronic obstructive pulmonary disease,¹⁵ diabetes¹⁶ and cancer.¹⁷ Another study also demonstrated that pharmacist-led anticoagulation management was associated with improved patient satisfaction and increased confidence in pharmacist oversight.³⁶ In this study, pharmacists’ education and counselling enhanced patients’ awareness of the importance of intensified INR monitoring and helped to improve their maintenance of a stable diet in the PST group, thereby enhancing TTR and reducing adverse events. Thus, we believe well-trained pharmacists would be capable of managing warfarin by using PST in China. Of note, studies from other countries indicated that PST may be a cost-effective intervention to improve TTR and patient satisfaction for anticoagulation management in eligible patients.^{37 38} However, several challenges need to be addressed prior to the implementation of PST in China. These include training a sufficient number of qualified pharmacists capable of delivering PST, ensuring that participating patients can properly use POCT devices, accurately report results and adhere to medication adjustments as instructed, as well as evaluating the pharmacoeconomic value of PST. We plan to use a Markov model to systematically evaluate the cost-effectiveness of PST in China in the near future.

Whereas, in some RCTs, no significant difference was observed in the incidence of adverse events between the PST and control groups. This can be attributed to two main factors. First, in these studies, both the control and PST groups received relatively high-quality anticoagulation, as TTR in the control group was above 60%^{30 31}; and second, the follow-up period in some studies was very short, for only 3 months,³² or the study population was small, with fewer than 150 patients in total.¹⁹

By using the binary logistic regression, we identified that receiving PST and younger age were associated with fewer adverse events. Previous studies have demonstrated that the incidence of bleeding in patients receiving warfarin therapy rises with advancing age.³⁹ The increased sensitivity to warfarin with advancing age may explain this finding.⁴⁰ Considering their higher likelihood of multiple comorbidities, frequent polypharmacy (increasing risks of warfarin–drug interactions) and the inconvenience of routine PT/INR monitoring, elderly patients on warfarin will be a key target population for PST model implementation in low- and middle-income countries.

Previous research has shown that medication adherence in chronic conditions tends to decrease as the duration of the disease increases.⁴¹ In this study, we showed that the total adverse event rate was 18.7% in UC group 2 compared with 15.2% in UC group 1. Although the difference was not statistically significant, the observed trend warrants further investigation into long-term adherence patterns. Effective measures are required to improve medication adherence for patients on long-term anticoagulation therapy. Pharmacist-delivered medication counselling and education may be an option and further studies are required to confirm it.

Several limitations should be noted. First, all healthcare institutions involved in the study were located in Hunan Province, China. To reduce bias, we included hospitals of different levels within the province, reflecting the diversity of hospital types in China. Second, our study population consisted only of adults. While self-testing has proven effective in paediatric populations,⁴² further research is needed to validate these results in China’s paediatric population. Third, this study was not able to blind patients or providers, who may have been influenced by the Hawthorne effect, as the additional attention in the PST group may itself have improved adherence. Finally, the intervention and high refusal rate during enrolment indicate a selected cohort of motivated, tech-comfortable patients, limiting generalisability to a broader population. The findings of this study should be interpreted with these limitations.

Conclusions

Compared with patients with infrequent testing in the UC group, the pharmacist-led PST with ongoing education and counselling led to better control of the quality of warfarin management. However, the application of pharmacist-led PST in China requires further validation through larger studies with more diverse populations, and the cost-effectiveness of applying this model in China also needs to be evaluated.

Data availability statement

No data are available.