Telemedicine increasingly serves as a primary point of entry into the health care system for patients, particularly in urgent care scenarios [1]. Physicians providing digital consultations are tasked with maximizing efficiency in terms of encounter duration while ensuring that prescriptions issued adhere to established guidelines, which is a crucial component of the cost-effectiveness and quality of telemedicine services [2].

The role of prescription decision support systems (PDSSs) is critical in the digital health care environment [3]. Within electronic health records (EHRs), PDSSs can streamline the amount of time that clinicians spend navigating complex medical terminology [4]. Research indicates that standardizing data input can enhance routine documentation in medical records [5]. Apart from time efficiency, EHRs have been linked to improved care quality and increased compliance with clinical guidelines [6]. Although EHR adoption is associated with significant challenges, the development of strategies to facilitate their use is gradually progressing [7]. To date, however, research exploring voluntary PDSS use and its impact on the outcomes of direct-to-consumer telemedicine consultations for acute conditions has been lacking.

We propose that physicians’ adoption of a PDSS will demonstrate both a reduction in the time needed to deliver care and increased adherence to clinical guidelines. This study explored the relationships between telemedicine use and ICD (International Classification of Diseases)-triggered PDSS scores, consultation duration, and guideline adherence.

Throughout the 6-month study period, we analyzed a total of 10,485 encounters with 9644 patients. Patient demographic data and the most common diagnoses are described in Table 2. These encounters resulted in 12,346 prescriptions being issued by 290 different attending physicians. Notably, some patients had multiple consultations within the study timeframe, and multiple prescriptions were occasionally dispensed during a single consultation; these prescriptions were not individualized. Before the implementation of the self-report prescription system, 4497 (36.42%) prescriptions were issued, which increased to 7849 (63.58%) after its implementation. A preliminary test of the system commenced in 2021; however, the system was not fully operational and made available for voluntary use by all physicians until January 2022. Following its implementation, the automated prescription feature was used in 5022 (40.67%) of the encounters.

During the brief trial period of the PDSS tool in 2021, a total of 261 (5.80%) prescriptions were made with electronic assistance, while 4236 (94.19%) prescriptions were issued without such assistance. Following the full deployment of the PDSS tool in 2022, the figures shifted significantly, with 5022 (63.98%) prescriptions being made with electronic assistance compared to 2827 (36.02%) without assistance.

The data demonstrated a significant reduction in consultation time after self-reporting: 9.5 (SD 5.5) minutes versus 11.2 (SD 5.9) minutes (P<.001; Mann-Whitney U test=12,118,181.5; Figure 1). Figure 2 shows that the decrease in the average consultation duration was correlated with service density, which was defined as the volume of services delivered by the center within a specified timeframe.

Regarding guideline adherence among the 12,346 medications prescribed, a substantial number of these medications (n=8683, 70.34%) were prescribed in accordance with guidelines. Notably, compared with manual prescription entry, the use of the automated filling system significantly increased guideline adherence (n=4697, 93.5% vs n=1389, 49.1%; P<.001; z score=45.24).

Regarding adherence to the institutional protocol for medical prescriptions, when the PDSS tool was not used, 1438 (50.87%) prescriptions deviated from the standard recommendations, while 1389 (49.13%) complied. Our analysis demonstrated that the application of the PDSS tool significantly improved adherence, with only 325 (6.47%) prescriptions failing to comply with the protocol, compared to 4697 (93.53%) that were aligned.

This study demonstrated that shortly after becoming available, most physicians opted to voluntarily use the decision support autocomplete functionality for medical prescriptions. Innovative approaches that facilitate the use of EHRs tend to be well received by medical practitioners [8]. The integration of new information technologies has transformed numerous aspects of health care and thus revolutionized care delivery within health systems [9], representing a significant advancement in digital health care, although certain challenges still need to be overcome. These challenges include poor interface designs, suboptimal performance, maintenance issues, overreliance, and dependency, all of which could jeopardize patient safety [10].

This study revealed an impressive rate of 94.7% adherence to recommendations, suggesting a potential improvement in the safety of telemedicine consultations with the use of autocomplete prescriptions. Guideline adherence is a measure of how well health care providers follow established clinical guidelines and protocols and is crucial for minimizing risks and ensuring optimal care quality and safety [11]. Notably, given that evidence suggests frequently deficient adherence with respect to in-person consultations, telemedicine shows promise as a method of care associated with better compliance [12-14].

The safety of prescribing practices and reductions in medication errors have been supported by evidence from electronic prescriptions selected through the use of standardized medication lists, codified instructions, and multimodal decision support [15]. Nonetheless, the prevalence of electronic prescribing errors can be high, with rates nearing 60%, particularly when considering incorrect field entries [16]. Previous studies have identified numerous potential error types in electronic prescriptions [17], and the electronic prescription strategy chosen has been noted to be a source of medication errors [18], which underscores the importance of continuing to improve electronic prescription systems.

The option for voluntary prescription autopopulation has several potential benefits for health care delivery. This approach may alleviate bureaucratic burdens associated with current medical practices, thereby enhancing communication with patients [19]. In high-demand settings, an autocomplete option can optimize consultation times as part of a broader management strategy [20]. Additionally, access to checklists for possible medication recommendations may reduce bias [21].

Our analysis revealed a 15.18% decrease in physicians’ average consultation time, from 1123 to 949 minutes, when the autocomplete feature was used, although reliance entirely on manual prescription entry was noted in 31% of consultations. Intriguingly, even with voluntary use, most prescriptions issued using the autocomplete function strictly adhered to medication recommendations, indicating high adherence to these recommendations. Prior to our study, no research had specifically investigated telemedicine quality based on guideline-directed prescribing facilitated by an autocomplete function. Accordingly, the establishment of high-quality telemedicine centers that continuously update management guidelines and develop strategies to meet policy requirements and deliver excellent remote medical consultations is imperative to enhance EHR systems. In addition to prescribing practices alone, administrators need to ensure the usability and appropriate implementation of coding for the autocomplete method to prevent misuse and maintain safety.

No specific security assessment was conducted. However, protocol adherence is recognized as a primary indicator of safety in care delivery, including referrals for in-person care when warning signs are identified. This study revealed enhanced adherence to guideline recommendations through the use of the self-report system, indirectly suggesting that increased safety was observed within this group.

Among the limitations of this study, the observed outcomes from the use of the clinical decision support system may not be solely attributable to changes in care practices. Instead, these outcomes might also reflect changes in the ICD codes selected by physicians, possibly for the convenience of prescribing. Moreover, the clinical support decision system may have encouraged excessive prescribing despite many of these conditions being manageable with nonpharmacological interventions, especially for those with mild symptoms. Notably, care quality assessed based solely on whether ICD codes matched medication recommendations may not effectively represent the true quality of care and may reflect decreased care quality, as previously mentioned. Another possible limitation was that the group that used decision support may have included patients with simpler diagnoses.

In this study, a PDSS was voluntarily used by physicians for most encounters. Decision support incorporation for prescription selection within EHRs, particularly in scenarios where policies are clearly established, may contribute to reducing consultation times and promoting high rates of adherence to guideline-directed prescriptions. This finding highlights the potential role of these systems in improving both the efficiency and quality of health care delivery, especially within telemedicine environments where prompt and precise decision-making is essential.