The study draws on data from a publicly reimbursed pilot program, implemented as part of a quasi-experimental, multisite trial in Estonia between January 2022 and July 2023, across 7 family medicine centers and 1 dermatovenerology outpatient clinic. This design enabled the integrated analysis of quantitative and qualitative data, providing a comprehensive understanding of the intervention’s effectiveness, feasibility, and real-world implementation. Mixed methods approaches are widely recognized as suitable for evaluating complex health services, particularly when both outcomes and adoption processes are key concerns [29-31].

To comprehensively evaluate both the clinical outcomes and contextual implementation of remote psoriasis monitoring, the study used a retrospective, convergent parallel mixed methods design. Quantitative data included patient-reported Dermatology Life Quality Index (DLQI) scores, clinician-reported Psoriasis Area and Severity Index (PASI) scores, and claims-based health care utilization data. Quantitative analyses, such as comparative effectiveness and correlation techniques, were conducted to measure changes in QoL and identify clinical or demographic factors influencing these outcomes.

To complement the quantitative findings, a qualitative case series was conducted to explore variations in care pathways and management practices at the primary care level. Although no formal thematic analysis was performed, the qualitative component provided illustrative, patient-level insights based on observational and contextual data. These data were drawn from pseudonymized patient feedback questionnaires (administered online via Google Forms; Google LLC/Alphabet Inc) and system usability assessments. This approach enhanced the interpretation of the findings by capturing care process differences and patterns that quantitative analysis alone could not reveal. It was also used to illustrate individual variability and highlight potential mechanisms of improvement.

As a result of feasibility and ethical constraints, patients were assigned to groups based on their level of care (primary or specialist) and willingness to participate. Randomization was not possible within the scope of this publicly funded pilot intervention.

Data sources were a nationwide public insurance claims database covering all publicly reimbursed patient interactions, visits, and medications during the study period, as well as a secure TD platform database where DLQI and other remote monitoring measures were reported.

While the broader pilot project evaluated multiple metrics (eg, adherence, cost, usability), this study specifically focuses on QoL outcomes (DLQI) and the comparative analysis of care models, using a subset of the intervention group data.

The primary outcome of this paper was the effectiveness of remote psoriasis management, measured by changes in DLQI scores across primary and specialist care settings [32,33]. The DLQI, a validated 10-item questionnaire, assesses the impact of psoriasis on QoL. Scores range from 0 to 30, with higher scores indicating greater impairment [34,35]. This tool is widely used in dermatological studies, enabling comparisons and providing insights into factors influencing outcomes [5,35,36].

Effectiveness was first analyzed for the entire intervention group, then compared between primary and specialist care, and finally evaluated separately to provide detailed insights. PASI scores at baseline were used to categorize psoriasis severity (mild: 1-4, moderate: 5-10, severe: >10) [37], offering clinical context for examining the relationship between disease severity and QoL outcomes.

Correlation analysis was conducted to explore clinical factors associated with changes in DLQI scores in primary care. The odds ratio for achieving a meaningful DLQI improvement (minimal clinically important difference [MCID]≤–3) was calculated, based on prior studies defining MCID ranges in dermatological conditions [38,39]. Specific cases that achieved the MCID in psoriasis management at the primary care level were selected for case series analysis, integrating qualitative and quantitative data such as patient feedback questionnaires, claims data [40,41], and medication prescriptions.

This study analyzed data from 110 patients participating in the remote monitoring group across both care levels in the real-life, publicly reimbursed remote monitoring pilot trial. The cohort consisted of 55 men and 55 women, with 84 (76.3%) aged 50 years or younger and 26 (23.6%) aged over 50 years. Inclusion criteria for remote monitoring were (1) a diagnosis of plaque psoriasis, (2) age between 18 and 75 years, (3) access to a smartphone or computer, and (4) willingness to provide informed consent and commit time to the study. Exclusion criteria were (1) a diagnosis other than plaque psoriasis, (2) a history of alcohol or drug abuse within 6 months before the study, (3) inability to fulfill study requirements, and (4) any other reasons deemed by the investigator or sponsor that rendered the patient ineligible.

Data were collected at both specialist and primary care levels. Patient eligibility was first confirmed, and informed consent was obtained, followed by the initial completion of patient-reported outcome measures [42] and PASI assessments, all securely stored in a specialized TD database developed by a TD company for this use case. The study protocol required patients to regularly complete online questionnaires such as the DLQI, Psoriasis Symptoms and Signs Diary [43], and Early Arthritis for Psoriatic Patients [44], facilitated through a web-based platform with a 2-factor authentication log-in. DLQI questionnaires were scheduled to be sent monthly on the 15th, allowing patients to complete them at home. Email reminders, including follow-up notifications, were also sent to encourage timely completion. Doctors were also encouraged to actively engage with patients. The platform enabled real-time transmission of patient-generated data, allowing physicians to securely and conveniently access and review the results. Both PCPs and specialist dermatologists could review this information asynchronously to provide recommendations, initiate treatment adjustments, or refer patients to other specialists as needed.

The System Usability Scale (SUS) was used to assess system usability through a 10-item questionnaire. Participants rated each item on a 5-point Likert scale ranging from 0 (strongly disagree) to 4 (strongly agree). Individual responses were converted and summed, then scaled by a factor of 2.5 to produce scores ranging from 0 to 100, where higher scores indicate better usability [45]. SUS scores were obtained directly from the pilot study without adjustments and were analyzed as part of the case series outcomes in this study.

Additionally, e-consultations were facilitated through the health information system. PCPs submitted detailed referrals to a chosen specialist, including information about the patient’s condition, test results, clinical observations, and secure access to previous remote monitoring data collected at the primary care level. Specialists reviewed the referral and provided recommendations or requested in-person visits, responding within 4 working days. Both the PCP and the patient had access to the consultation records, fostering transparency and efficient communication [46].

The intervention was financed through a public reimbursement model implemented as part of a nationwide remote care pilot project by the Estonian Health Insurance Fund, the sole public payer in the solidarity health care system. This ensured that TD services were accessible to the population without additional costs to the patient [47]. The payment model incorporated value-based elements, including outcome-based payments to participating health care providers at the end of the 1-year remote monitoring period.

Participants were enrolled at different times throughout the study period. To account for this, the first DLQI entry within 45 days of enrollment was considered the baseline score. If the QoL questionnaire was completed twice during the initial or any follow-up month, the average of the 2 scores was used. Each participant’s follow-up timeline was anchored to their individual start date, with the first month of follow-up defined according to whether enrollment occurred before or after the 15th of the calendar month. This approach ensured alignment across patients despite rolling enrollment.

Participants were excluded from the study if they had more than 6 months of inactivity or if they did not complete a baseline DLQI questionnaire within 45 days of enrollment. Participants meeting both exclusion criteria were also excluded from the final analysis.

Patient satisfaction was measured using a feedback form adapted from the Estonian Society of Family Physicians’ standard questionnaire, with detailed information provided in Multimedia Appendix 1. The questionnaire was administered during the final quarter of the pilot project via Google Forms. Responses were collected online, pseudonymized by the research team, and then linked to individual patients in the study database for integration with quantitative data.

The study received ethical approval from the Research Ethics Committee of the University of Tartu, Estonia (ethical approval code 350/T-14). All participants provided written informed consent before enrollment, either in person or electronically, after receiving detailed information about the study and having the opportunity to ask questions.

Data were collected from routine clinical care and study-specific instruments (e.g., questionnaires and photographs). Identifiable data were pseudonymized or anonymized before analysis. Data were securely stored in REDCap (Vanderbilt University; hosted by the University of Tartu) and in a secure teledermatology (TD) platform database, which supported digital patient communication and image handling. Access to both systems was restricted to authorized personnel only.

Pseudonymized data and identification keys will be stored until December 31, 2030, after which the keys will be destroyed. Fully anonymized data will be kept until December 31, 2050, to support future research.

Participation in the study was voluntary, and no compensation was provided to participants.

Of the 110 patients enrolled in the intervention, 103 provided DLQI scores. Of these, 76 met the criteria for inclusion in the analysis, with a balanced distribution of males (n=37, 49%) and females (n=39, 51%). The mean age of participants was 43.46 (SD 10.86) years, spanning early to middle adulthood. The mean baseline PASI score was 3.94 (SD 4.46), indicating mild to moderate psoriasis on average. Participants’ baseline DLQI scores averaged 5.28 (SD 4.90), reflecting a moderate impact of the condition on their QoL. Additionally, 27 (36%) participants reported having comorbid conditions, including hypertension (n=8, 30%), dyslipidemias (n=5, 19%), and enteropathic arthropathies (n=4, 15%), among others. The mean duration of psoriasis since onset was 11.70 years (SD 6.12 years), indicating that many participants had lived with the condition for a substantial part of their lives. All results are presented in Table 1.

A comparison of baseline characteristics between included (n=76) and excluded (n=27) patients was conducted to assess the risk of selection bias resulting from the exclusion criteria. Of the excluded patients, 15 were from primary care and 12 from specialist care.

No statistically significant differences were observed in any of the assessed variables, including age (P=.15), sex (P>.99), baseline DLQI (P=.96), PASI score (P=.26), years since onset (P=.82), or clinic type (P=.54), indicating that exclusions did not systematically bias the final analytic sample. Full comparison results are provided in Multimedia Appendix 3.

In the remote monitoring group, DLQI scores decreased by an average of –0.85 (SD 2.89), indicating a general improvement in QoL across the combined care group. The Wilcoxon signed rank test confirmed that this mean change was statistically significant (mean difference –0.85, 95% CI –1.48 to –0.23, W=823, P=.009).

In the specialist care group, the mean DLQI change was –1.33 (SD 3.12), reflecting a statistically significant improvement (mean difference –1.33, 95% CI –2.34 to –0.32, t₃₈=–2.66, P=.01). By contrast, the primary care group showed a mean DLQI change of –0.34 (SD 2.57), which was not statistically significant (mean difference –0.34, 95% CI –1.00 to 0.42, W=259, P=.36). Detailed results are presented in Tables 2 and 3.

In a post hoc linear regression model incorporating enrollment month, clinic type, age, sex, and e-consultation frequency as covariates, no significant association was found between enrollment timing and DLQI change (β=–.025, P=.92), indicating that staggered enrollment did not introduce meaningful bias. Full model coefficients are provided in Multimedia Appendix 4.

An independent samples t test was conducted to compare the mean DLQI change scores between patients receiving care in specialist settings and those in primary care. The difference was not statistically significant (mean difference 0.99, 95% CI –0.32 to 2.30, t₇₄=1.51, P=.14). The Levene test confirmed the assumption of equal variances (F_1,74=2.12, P=.15).

The analysis included patients receiving specialist care (n=39) and primary care (n=37). A clinically meaningful improvement in DLQI—defined as a change of ≤–3—was achieved by 10 patients in the specialist care group and 3 patients in the primary care group.

Statistical analysis revealed a significant difference in MCID achievement between care settings (χ²₁=4.12, P=.04). Patients receiving specialist care were approximately 3.91 times more likely to attain a clinically meaningful improvement in DLQI compared with those in primary care (see Table 4).

Analysis within the primary care group showed no significant correlation between baseline PASI scores and DLQI change (r=0.15, P=.38). However, a moderate negative correlation was found between the number of e-consultations—defined here as collaborative care interactions between PCPs and specialists—and DLQI change (r=–0.34, P=.04). This suggests that increased e-consultations were associated with greater improvements in QoL, reflected by larger decreases in DLQI scores. The key findings are summarized in Table 5.

As a mixed methods study, it balanced quantitative and qualitative data by integrating several complementary research methods, which introduced complexities in design and execution and increased the risk of bias [66].

The relatively small sample size (n=76) limits the generalizability of the findings from the comparative effectiveness and correlation analyses. A larger sample would provide more robust data and enable more definitive conclusions about differences in QoL outcomes between specialist and primary care settings. This study did not perform an a priori sample size calculation, as it was based on data from a real-world pilot intervention [67]. The final sample size was determined by pragmatic and logistical constraints, which may reduce the statistical power of subgroup analyses and the overall generalizability of the findings. Additionally, the study lacked control for confounding factors such as variations in psoriasis severity and treatment adherence.

Although the pilot design did not explicitly control for time-varying external factors, a post hoc analysis incorporating enrollment month found no significant association with DLQI change, indicating limited temporal confounding due to staggered recruitment.

While DLQI improvements were observed, especially in the specialist group, these may partially reflect improved patient adherence or physician-led treatment adjustments in response to remote monitoring data. Although the pilot aimed for patients to complete at least one patient-reported outcome measure each month, several participants submitted DLQI responses only a few times. This limits our ability to assess the long-term effectiveness and sustainability of remote care models based on DLQI data alone. While SUS scores were collected from a broader sample, analysis was limited to the 3 primary care patients included in the case series. Therefore, usability-related findings should be interpreted as illustrative and are not generalizable to the full cohort. Broader usability and access challenges, such as internet connectivity issues, device limitations, or low digital literacy, were not measured but could have influenced patient engagement and outcomes. These factors should be considered in future studies evaluating digital care models.

Although medication and prescription data were available for all participants, a detailed review was conducted only for the 3 primary care patients who achieved an MCID≤–3 in DLQI. This targeted review aimed to explore potential factors associated with clinically meaningful improvement in primary care settings, where overall outcomes were less favorable compared with specialist care. By focusing on patients who demonstrated significant QoL improvement, we sought to identify possible contributors, such as treatment changes, collaborative care episodes, or patient engagement, that may explain better outcomes in the absence of specialist-led management.

These exploratory insights highlighted variability in treatment pathways [30] at the primary care level. However, future research should systematically analyze medication use across the full cohort to better understand the relationship between treatment changes and patient-reported outcomes, and to distinguish these effects from other potential contributors, such as usability, communication, or perceived support.

In addition, comparisons of baseline characteristics between included and excluded patients revealed no statistically significant differences, suggesting a minimal risk of selection bias in the final analytic sample.

The study included a case series analysis to identify factors and patterns. This analysis did not reveal a common pattern contributing to QoL improvement among patients with psoriasis in primary care remote monitoring. However, it also did not contradict the potential factors identified in existing literature and supported by our correlation analysis [57]. The remote monitoring pilot did not prescribe clear protocols for collaborative care or referral practices. Therefore, these factors were investigated using quantitative public claims data alongside less rigorous exploratory methods.

Variability in remote care management approaches among PCPs, coupled with the absence of clear collaborative care protocols, may have limited the consistency and rigor of outcomes observed in this study. Additionally, the study may be subject to selection bias, as participants were recruited from different sites and levels of care. Differences in care delivery models, such as PCP-led versus nurse-led workflows, could have introduced performance bias, potentially affecting the consistency of outcomes.

Based on this, it is clear that any remote monitoring study conducted across different care levels (primary and specialist) should rigorously consider the impact of the health care system’s referral practices on QoL, such as whether general referral, e-consultation, or collaborative care elements are already integrated into the system. These factors should be accounted for, and protocols for using such referral practices should be clearly outlined for study participants.

This is the first study to compare publicly reimbursed remote psoriasis monitoring at both specialist and primary care levels while also evaluating collaborative care tools such as e-consultations and standardized referral protocols aimed at improving patient outcomes. Its limitations provide a foundation for future research, which should address these gaps through well-powered randomized controlled trials that incorporate the factors and insights identified in this study.

Future studies should prioritize developing structured collaborative care models, such as standardized e-consultations, to enhance PCP-led psoriasis management. Another promising research avenue is leveraging public claims data to analyze care processes and identify patterns in referral and treatment behaviors. As noted in prior studies, investigating the psychological effects of remote monitoring could offer valuable insights into how these interventions influence patient security and reassurance.

Moreover, future research should investigate variations in care design and compare the effectiveness of nurse-led, PCP-led, and collaborative approaches. Well-powered randomized controlled trials with larger and more diverse populations are essential to validate findings and enhance our understanding of how remote monitoring can improve QoL outcomes, particularly in primary care settings.

With limited access to specialists, PCPs play a crucial role in managing psoriasis, underscoring the importance of equipping them with appropriate tools and support. Collaborative practices, such as e-consultations and standardized referral protocols, can enhance care coordination and complement specialist expertise. Providing PCPs with improved resources and guidance can contribute to better patient outcomes, especially in resource-constrained settings.

This study provides unique insights by directly comparing remote monitoring outcomes between specialist and primary care settings. It demonstrates that primary care, when supported by structured resources, can approach specialist-level effectiveness. However, remote management in primary care results in less significant improvements, reflecting systemic challenges also observed in in-person PCP-led care.

These findings highlight the potential of integrating specialist involvement into PCP-led care through standardized e-consultation practices. Leveraging public claims data within a mixed methods approach underscores the need for collaborative care models to ensure consistent quality. Further research should focus on refining and implementing these models to optimize remote psoriasis management in primary care.