The prevalence and disease burden of chronic health conditions is on the rise. The World Health Organization predicts that by 2030, chronic noncommunicable health conditions will account for >50% of the total disease burden [1,2]. In particular, cardiovascular conditions, cancer, respiratory conditions, and diabetes have the highest morbidity and mortality [1]. Currently, 60% of the US population has at least 1 chronic condition and 42% of the population has multiple chronic conditions [3]. This results in a high individual disease burden owing to the large impact on social participation and required patient self-management skills. Self-management refers to a person’s ability to manage the clinical, psychosocial, and societal aspects of their illness and its care [4]. In contrast, self-efficacy is a person’s belief that he or she can successfully execute this behavior [4]. Apart from a high individual disease burden, the prevalence of chronic conditions imposes a high macroeconomic burden [5]. Furthermore, an increasing shortage of health care providers is expected, among others in the United States [6] and Europe [7,8]. In combination with the increased pressure put on health systems by unexpected events such as the COVID-19 pandemic, this shortage threatens the delivery of essential health services [9]. To preserve the access to care for all patients, new technologies are increasingly being developed and adopted, including patient-centered digital health records.

Such patient-centered digital health records can significantly help engage and empower patients with a chronic health condition [10-13]. Patient-centered digital health records enable patients to take on a more active role in their care by allowing them to view parts of their medical records, such as medication lists, laboratory and imaging results, allergies, and correspondence. Other common features include secure messaging, requesting prescription refills, video consultation, paying bills, and managing appointments. Examples of patient-centered digital health records include patient portals and personal health records (PHRs). Patient-centered digital health records differ in the volume and detail of the provided medical data, functionalities, and level of patient control, as shown in Textbox 1. Highly engaged patients are reported to have increased disease knowledge, better self-management, more self-efficacy, and improved clinical outcomes [14-16]. The effects of using patient-centered digital health records may be most substantial for patients with chronic conditions. Many self-management skills are required, and their potential gains are the highest. Not only patients but the entire health care system might benefit from an increased adoption of patient-centered digital health records.

Currently, huge investments of time and resources are made in patient-centered digital health records. However, limited insight exists in how the use of patient-centered digital health records by patients with a broad range of chronic conditions affects clinical and patient-reported outcomes and health care utilization. Moreover, we lack an overview of their effects on patient satisfaction, and the feasibility and acceptability of their use by people with chronic conditions. Previous systematic reviews focused on one health condition [23], focused on one type of digital health record [24-27], investigated a select set of health outcomes [24,26,28], or are now obsolete in this rapidly changing technological landscape [23,25,27].

Therefore, in this systematic review, we summarized the available evidence on patient-centered digital health records. Our primary objective was to assess how patient-centered digital health records for nonhospitalized patients with chronic conditions affect clinical and patient-reported health outcomes and health care utilization. Our secondary objective was to evaluate patient satisfaction with and feasibility and acceptability of using patient-centered digital health records. Results of this systematic review may help guide future development and implementation.

A medical librarian (MB) conducted the original literature search using the following databases: MEDLINE, Cochrane Library, CINAHL, Embase, and PsycINFO. All original studies published between January 1, 2000, and December 1, 2020, were assessed. A search update in MEDLINE was performed for all studies published between December 1, 2020, and December 31, 2021. Multimedia Appendix 1 presents the full search strategy. Articles published before 2000 were excluded because of the rapidly changing field of digital health technology [30].

Patient-centered digital health records were defined as mobile health (mHealth) or eHealth technologies that enable a 2-way exchange of health information between patients and the health care system, such as patient portals, PHRs, or mHealth apps with a health record functionality. A patient-centered digital health record provides patients with the ability to view, download, or transmit their health information on the web. This health information was updated at regular intervals. In addition, a patient-centered digital health record allows for communication between patients and the health care system, either by adding or editing health information, exchanging patient-reported outcomes, or by using communication tools such as messaging. Several other functionalities are common, but were not considered essential; for example, appointment scheduling, requesting prescription refill, viewing educational material, using decision support tools, and using connected wearables. Exclusion criteria were nondigital health records, digital health records intended for hospitalized patients, and digital health records that are not accessible to patients, such as the clinician-facing components of the electronic health record (EHR).

Studies investigating patient-centered digital health records intended for nonhospitalized patients with a chronic health condition were included. Only studies published in English were included. Eligible studies included randomized controlled trials (RCTs), quasi-experimental studies, nonexperimental observational studies (including cohort and cross-sectional studies), and pilot or feasibility studies. Of mixed methods studies, only nonqualitative parts were used for data extraction. Studies that only described health care providers’ experiences were excluded.

Studies on patients with a chronic health condition of all age groups were considered. Chronic conditions included all diseases with a moderate to high disease burden and moderate to high impact on daily life. Consequently, these conditions demand considerable self-management skills from patients to manage the clinical, psychosocial, and societal aspects of chronic condition and its care. The selection of chronic conditions included in our search strategy was based on the Charlson Comorbidity Index, other literature, and clinical expertise [31,32]. Diseases included cancer, arthritis, HIV, AIDS, asthma, chronic obstructive pulmonary disease, chronic heart conditions, hematologic disease, chronic kidney disease, celiac disease, inflammatory bowel disease, cystic fibrosis, diabetes mellitus, and multiple sclerosis (MS).

Studies were required to report at least one primary or secondary outcome. Primary outcomes were clinical outcomes (including disease events and complications, vital parameters, and laboratory parameters), patient-reported outcomes (including self-management and self-efficacy, patient engagement, health-related quality of life (HRQoL), stress and anxiety, and treatment adherence), and health care utilization (including the number of emergency department [ED] visits and hospitalizations, the use of preventive or recommended care services by patients, and regular workload for health care professionals). Secondary outcomes included technology-related outcomes (including patient satisfaction, feasibility, and acceptability). Definitions and examples of these 13 outcomes are presented in Table 1.

Two independent reviewers (MB and SB) assessed titles, abstracts, and full texts for eligibility. Disagreements were resolved by discussion, if necessary, with a third reviewer (SG).

A modified, electronic version of the standardized Cochrane data extraction form [39] was used to extract the following data items: first author’s name; publication year; study design; disease or diseases studied; study aim; country and setting; participants’ age and sex; sample size; inclusion and exclusion criteria; follow-up duration; description, features, and purpose of the patient-centered digital health record and (if applicable) of the comparator; size and description of the control group (if applicable); device used; description of health outcomes and results; and main study findings.

For quality appraisal, Joanna Briggs Institute (JBI) critical appraisal tools for RCTs, cross-sectional studies, cohort studies, and quasi-experimental studies were used [40]. JBI tools were modified to better suit the assessment of digital health record studies. Several items were added, including adequate patient-centered digital health record descriptions and selection bias measures, as presented in Multimedia Appendix 2. As the JBI tools differed in the number of items, all scores were converted to a 15-point scale. Articles with a score of ³12 were considered of “high quality,” between 8.5 and 11.9 of “medium quality,” and <8.5 of “low quality.”

Associations between patient-centered digital health record use and health outcomes were categorized in 3 groups: “beneficial,” “neutral or clinically nonrelevant,” or “undesired.” Categorizations were determined by our interpretation of study findings, based on meaningful clinical effects and statistical significance (P<.05), and could therefore differ from the authors’ conclusions. Statistical significance was considered relevant only if the effect size were clinically significant. If available, minimal clinically important differences were used to assess effect sizes. The summarization of effects was based on the vote-counting method, as no meta-analysis could be performed. The findings were summarized for all conditions, grouped by disease category (diabetes mellitus, cardiopulmonary diseases, hematology-oncology diseases, and other diseases), and grouped according to outcome type (clinical outcomes, patient-reported outcomes, health care utilization, and technology-related outcomes).

Several subgroup analyses were performed. The first subgroup included conditions with a high disease burden. These included conditions with either impaired social participation or that require a high level of self-management skills. Impaired social participation was defined as being unable to participate in work or school or engage with friends and family as desired because of the condition or its treatment. High self-management skills are defined as recurrent actions demanded from patients to prevent or treat the disease or its consequences, including high disease-related knowledge needed to actively engage in decision-making. This subgroup was determined based on clinical expertise of the study team. Second, we assessed 2 subgroups: patient-centered digital health records that predominantly offered passive features and those that predominantly offered active features. Passive features are those through which the patient receives information but does not actively add information. Active features are those in which the patient performs an action and actively engages with the digital health record. The third subgroup of interest included studies with high methodological quality. A sensitivity analysis was performed to investigate whether our results were influenced by poor quality studies. Finally, the subgroups of interest were studies that included older participants (mean age >55 years), a high number of female participants (>45%), or a racially diverse population (<50% White participants).

The search yielded 7716 unique publications. After screening the titles and abstracts, 320 full-text articles were retrieved. A total of 81 articles met the inclusion criteria. No non-English articles that met the inclusion criteria were identified. Figure 1 shows the study PRISMA flowchart. In total, 1,639,556 participants were included in the studies of this systematic review. Most (74/81, 91%) studies included only adult participants. Of the total 1,369,913 participants, 99% (n=1,629,660) were adults. Nine studies included children or their parents, with a total number of 9297 children and 599 parents. Sample sizes of studies varied from 10 to 267,208 participants. Furthermore, 46% (747,370/1,639,556) of the participants were female. Of the 81 included studies, health literacy was reported by 7 (9%) studies and insurance status by 15 (20%) studies. Race distribution was reported by 74% (60/81) of studies, of which 47 (78%) studies included a population of which more than half were White and 26 (43%) studies of which >75% were White.

Study characteristics are presented in Tables 2-5 (36 studies are listed in Table 2; 11 studies are listed in Table 3, 14 studies are listed in Table 4, and 20 studies are listed in Table 5). Most investigated conditions were type 1 or 2 diabetes mellitus (37/81, 46%), cardiovascular conditions (14/81, 17%), and malignancies (11/81, 14%). Studies were mostly conducted in the following countries: United States (58/81, 72%), the Netherlands (7/81, 9%), Canada (5/81, 6%), and United Kingdom (3/81, 4%). In addition, 30% (24/81) of the studies were RCTs, 27% (22/81) were cross-sectional studies, 20% (16/81) were retrospective observational cohort studies, and 23% (18/81) were quasi-experimental studies, including pretest-posttest and feasibility studies. One study was a secondary data analysis of the intervention group in an RCT. Of the 55 studies that reported follow-up durations, 6 (7%) studies had a follow-up of less than a month, 25 (31%) studies between 1 and 6 months, 14 (17%) studied between 7 and 12 months, and 10 (12%) studies of >12 months.

Explanations of the patient-centered digital health records investigated in each study are presented in Tables 6-9. Patient-centered digital health records range from a pilot patient portal enabling patients to view a limited set of their medical data to comprehensive PHRs, offering extensive data access and enabling appointment scheduling and prescription refill requests. A minority (12/81, 15%) of studies specifically evaluated ≥1 digital health record features such as secure messaging or a medication adherence module. In addition, 15% (12/81) of studies used a hybrid approach to assess a combination of a digital health record with a connected device, or with training, coaching, or face-to-face visits.

An overview of reported associations for each health outcome is shown in Figure 2. The proportions of beneficial effects reported per health outcome are presented in Multimedia Appendices 3 and 4. For high-quality studies, proportions are presented in Multimedia Appendix 3. An overview of study conclusions and associated outcomes is presented in Tables 10-13. Studies were grouped according to disease group.

In 44 studies investigating a total of 69 clinical outcomes, a beneficial association with digital health record use was reported for 42% (29/69) of the outcomes. Hospitalizations and exacerbations were the most frequently studied disease events and complications, with beneficial effects reported in half of the studies (2/4 and 2/4, respectively). Blood pressure was the most frequently studied vital parameter, with beneficial effects reported in 36% (5/14) of the studies. HbA_1c and cholesterol levels were the most frequently studied laboratory parameters, with beneficial effects reported in 53% (10/19) and 57% (4/7) of the studies, respectively. No clinical outcomes were unfavorably affected by patient-centered digital health record use. In comparison with the total population, higher proportions of beneficial effects were reported for diabetes mellitus and cardiopulmonary diseases. When focusing on 14 high-quality studies, beneficial effects were observed less frequently, in only 30% (7/23) of the clinical outcomes.

Studies that assessed vital parameters generally reported few other health outcomes. However, among the studies that assessed disease events and complications, and laboratory parameters, beneficial effects were often associated with improved treatment adherence [52,68,71,81]. We hypothesize that this might be related to the removal of logistical barriers for patients in obtaining web-based prescription refills, as opposed to having to call health care providers or send them an email. Of the 6 high-quality studies that investigated treatment adherence, 2 studies assessed patient-centered digital health records that enabled patients to request prescription refills and found beneficial effects on adherence [52,68].

Overall, in 53 studies investigating a total of 86 patient-reported outcomes, a beneficial association with digital health record use was reported for 45% (39/86) of the outcomes. Of the 18 studies investigating 19 self-management or self-efficacy outcomes, beneficial effects were reported in 53% (9/19). Of these 9 studies, 56% (5/9) used validated questionnaires. For patient engagement outcomes, large differences in the proportions of beneficial effects were observed: from 11% (1/9) for patient activation, to 56% (5/9) for patient involvement, and 70% (7/10) for disease knowledge. However, only in measuring patient activation, validated questionnaires were principally used (8/9, 88% of studies). For HRQoL, beneficial effects were reported in 27% (4/15) of the studies, of which half used validated HRQoL questionnaires. No patient-reported outcomes were unfavorably affected by patient-centered digital health record use. In comparison to the total population, higher proportions of beneficial effects were reported for diabetes mellitus, especially for patient engagement and treatment adherence. Lowest proportions were reported for cardiopulmonary diseases, especially for patient engagement. When focusing on 10 high-quality studies, a lower proportion (7/19, 37%) of beneficial effects was observed.

We observed that improvements in patient engagement were especially facilitated by strengthening patient-professional communication; for example, through secure messaging [71,81,93]. In addition, both self-efficacy and HRQoL primarily seemed to be reinforced through the use of 2 functionalities: patient-professional communication [54,90,113] and information on disease progression [90,113].

For 24 studies investigating a total of 27 health care utilization outcomes, a beneficial association with digital health record use was observed for 59% (16/27) of the outcomes. The highest proportion (10/13, 77%) of beneficial effects was reported for an increased use of recommended care services. Of these 13 studies, 5 (38%) focused on recommended care services for people with uncontrolled disease, 4 (31%) on the use of preventive care services, and 4 (31%) on medical follow-up rates. In 25% (3/12) of the studies that assessed reductions in ED visits and hospitalizations, these were accompanied by an increased use of other care services, including outpatient clinic appointments and secure messaging. Compared with the total population, highest proportions of beneficial effects were reported for diabetes mellitus and hematological and oncological diseases. When focusing on 7 high-quality studies, lower proportions (3/9, 33%) of beneficial effects were observed.

For 39 studies investigating a total of 75 technology-related outcomes, a beneficial association with digital health record use was observed for 88% (66/75) of the outcomes. All (22/22, 100%) studies reported high patient satisfaction with accessing and using digital health records. Furthermore, 75% (6/8) of the studies reported high patient satisfaction with the effects of using digital health records. High feasibility was reported by 79% (15/19) of the studies, and high acceptability by 88% (23/26) of the studies. Highest feasibility was reported for digital health records intended for people with hematological and oncological diseases. Lowest feasibility and acceptability were reported for digital health records intended for people with cardiopulmonary diseases. When focusing on 6 high-quality studies, proportions of studies that found beneficial effects were similar.

A subgroup of 47 studies that investigated patients with a high disease burden or high self-management was assessed. The following conditions were included: malignancies (11 studies), asthma (9 studies), HIV infection and AIDS (6 studies), hematologic conditions (5 studies), chronic kidney disease (3 studies), chronic heart failure (4 studies), mental disorders (3 studies), multiple sclerosis (2 studies), inflammatory bowel disease (2 studies), rheumatologic conditions (2 studies), insulin-dependent diabetes mellitus (2 studies), atrial fibrillation (1 study), cystic fibrosis (1 study), and posttransplant patients (1 study). In general, the digital health records assessed in this subgroup were more often tailored to specific patient populations through the addition of specialized functionalities or connected wearables.

In comparison with studies investigating patients with no high disease burden, studies investigating patients with a high disease burden reported considerably higher proportions of beneficial effects for vital parameters, patient engagement, reductions in ED visits and hospitalizations, and for all technology-related outcomes. Considerably lower proportions of beneficial effects were reported for laboratory parameters, health-related quality of life, treatment adherence, and increased use of recommended care services. For the 9 high methodological quality studies on high disease burden or self-management, the proportions of studies that found beneficial effects were roughly similar.

Of the 81 studies, 41 (51%) of the studied patient-centered digital health records focused on passive features and 40 (49%) focused on active features. In comparison with digital health records with an active focus, more beneficial effects were observed among digital health records with a passive focus for laboratory parameters (9/16, 56% vs 7/17, 41%), self-management and self-efficacy (7/11, 64% vs 3/8, 38%), patient engagement (9/15, 60% vs 4/13, 31%), and for an increased use of recommended care services (5/6, 83% vs 5/7, 71%). Compared with digital health records with a passive focus, more beneficial effects were observed among digital health records with an active focus on disease events or complications (4/10, 40% vs 1/5, 20%) and reductions in ED visits and hospitalizations (4/6, 67% vs 1/6, 17%). However, when focusing on high-quality studies, higher proportions of beneficial effects were seen for digital health records with an active focus on all clinical outcomes, patient-reported outcomes, reductions in ED visits and hospitalizations, patient satisfaction, and acceptability.

Of the 81 included studies, 27 (33%) studies were graded as low quality, 38 (47%) as medium quality, and 16 (20%) as high quality (Tables 10-13). Studies investigating cardiopulmonary conditions were of the highest quality, with 29% (6/21) of the studies graded as high quality. Of the 24 included RCTs, 7 (29%) were of high quality. Only 38% (9/24) of the RCTs concealed allocation to treatment groups, and 67% (16/24) used intention-to-treat analyses. Of the 57 studies with other designs, 9 (16%) were graded as high quality. Overall, 15% (12/81) of studies reported power calculations.

Among the 65 studies that were graded as medium or low quality, only 35% (23/65) used reliable or validated tools for the measurement of all their outcomes and 48% (31/65) for part of their outcomes. Of these 65 studies, 10 (15%) studies took adequate measures to limit selection bias and 17 (26%) studies used a control group or randomized participants.

When focusing on the 16 high-quality studies, 3 functionalities appeared to be the most effective: secure messaging to lower barriers in patient-professional interaction, prescription refill functions to improve medication adherence, and information provision on disease progression. In addition, in 16 high-quality studies, the proportions of beneficial effects were similar for a subgroup of studies that included older participants (mean age >55 years), which included a high number of female participants (>45%), or included a racially diverse population (<50% White participants), as compared with the total population.

In this systematic review, we evaluated evidence on the effects of the use of patient-centered digital health records in nonhospitalized patients with chronic health conditions on clinical and patient-reported outcomes, health care utilization, and technology-related outcomes. Beneficial effects were most frequently reported for the use of recommended care services (10/13, 77%) and for 4 patient-reported outcomes: disease knowledge (7/10, 70%), patient involvement (5/9, 56%), treatment adherence (10/18, 56%), and self-management and self-efficacy (10/19, 53%). Regarding clinical outcomes, beneficial effects were reported in 42% (29/69) of the studies. Beneficial effects were least frequently reported for disease events and complications (5/15, 33%) and health-related quality of life (4/15, 27%). For digital health records that predominantly focused on active features, higher proportions of beneficial effects on nearly all health outcomes were observed among the high-quality studies.

In this study, we observed that patient-centered digital health record use may be associated with an increased use of recommended care services. Beneficial effects on ED visits and hospitalizations were mainly observed when accompanied by an increased rate of follow-up appointments or secure messaging [60,89,93]. This might imply that reducing ED visits and hospitalizations is primarily achieved by facilitating patient-professional communication.

Beneficial effects were most often reported for patients with diabetes or cardiopulmonary disorders. We suggest 2 explanations. First, the focus of digital health records has been directed toward patients with diabetes and asthma for some time because of the sheer number of people with these conditions. This could have resulted in higher-quality patient-centered digital health records and patients who were more accustomed to their use. Second, the relative improvements in health outcomes might be smaller among patients with a condition with a high disease burden because of a higher baseline level of self-management skills and disease knowledge.

The proportions of beneficial effects varied considerably between health outcomes, which may be explained by 2 reasons. First, outcomes with a higher proportion of beneficial effects were more often the primary study outcomes than the secondary outcomes. Digital health records were more frequently tailored for these outcomes, yielding higher beneficial effects. Second, outcome assessment was generally less robust for outcomes with a higher proportion of beneficial effects, such as self-management and patient engagement, which might have resulted in more false-positive effects.

Our results are more positive than those of the previous systematic reviews. This might be because of the increasing acceptance of digital health records, their improving quality, the increasing body of literature, or variations in digital health record definitions used. Two previous reviews found mixed effects on the use of portals on health outcomes and health care utilization [27] and reported positive effects on qualitatively assessed self-management in only one-third of the studies [25]. A recent systematic review that focused on portals intended for hospitalized patients found mixed results for patient engagement [26]. A systematic review that included only qualitative studies found that portal use was associated with positive effects on self-efficacy, treatment adherence, and disease knowledge [28]. In a review on eHealth interventions that aim to promote medication use, a weak association between digital health record use and health-related quality of life was observed [10]. This implies that digital health record engagement is not yet sufficient to affect patients’ overall health-related quality of life.

This systematic review has several strengths. Our search strategy was comprehensive, to account for the lack of consensus in digital health record terminology. In addition, a wide variety of health outcomes were considered relevant to determine the impact of digital health record use. However, several limitations of this study must be considered. First, comparisons between studies were difficult because of the variety in evaluated functionalities. A similar diversity was observed among the reported follow-up durations, participants’ ages, study sample sizes, and outcomes. Second, because it was not possible to perform a meta-analysis owing to the heterogeneity in reported (disease-specific) outcome measurements and effects, we used the vote-counting method. Therefore, we could not report the effect estimates and indicated directions of effects [122]. Third, owing to a lack of agreement on feasibility and acceptability thresholds, much is left to the authors’ discretion. Fourth, JBI critical appraisal tools rank every item equally despite being not equally important. Finally, publication bias could have resulted in overestimation of the positive effects of patient-centered digital health records. More studies with positive results have been published. In addition, many of the included studies assessed more “mature” patient-centered digital health records, which could have overestimated the effects.

We observed that high patient satisfaction rates did not fully reflect in other health outcomes. This can be partly attributed to acquiescence bias and satisficing [123]. Moreover, satisfaction was often reduced to a narrow ease-of-use questionnaire, instead of satisfaction with the contribution to overall disease management. Finally, several studies only included recurrent users in their analyses, which could falsely increase feasibility. Moreover, these recurrent users likely experienced positive effects of using digital health records, which would have resulted in an overestimation of effects in randomized studies with no intention-to-treat analysis and in all nonrandomized studies.

The voluntary adoption of patient-centered digital health records by patients might reflect an intrinsic, preexisting motivation for self-management and care engagement bias, which may overestimate their effects. Patient-centered digital health record use could even be considered a surrogate measure for engagement [109,124,125]. Thus, it might be best to consider digital health records as vehicles for empowerment, strengthening existing self-management capabilities [126,127].

The effects of using patient-centered digital health records on health outcomes are not always direct but often depend on intermediate steps. For example, requesting prescription refills might depend on the actions performed by (slow-responding) physicians, nurses, or pharmacies. Thus, if using a digital health record would have no observable effects on health outcomes, this could also be a result of these intermediate steps or unforeseen processes and may not be attributable to the use of the patient-centered digital health record.

The proportion of beneficial effects reported in high-quality studies was lower as compared with all included studies for clinical outcomes (30% vs 42%), patient-reported outcomes (37% vs 45%), and health care utilization (33% vs 59%). Nevertheless, the proportions are clinically relevant and promising considering this newly emerging field. The observed differences might be related to 4 factors. First, the selection of motivated, well-educated, digitally minded participants might have overestimated the results in most low- and moderate-quality studies. Second, most studies did not measure ongoing user activity, and assumed that registered users became recurrent users. Third, nearly all low- and moderate-quality studies reported high dropout rates, which could overestimate acceptance rates. Finally, the lack of consensus on digital health record terminology hindered the interpretation of findings. We would advocate the use of uniform definitions, such as those presented in Textbox 1 [10,17-20].

Future studies should adopt additional measures to adhere to a uniform taxonomy, use log data, and limit selection bias. The exclusion of less-engaged people could further expand the digital divide between patients who are digitally proficient and those who are not, resulting in an increasingly unequal distribution of care services. We suggest that researchers include a diverse population based on age, gender, disease burden, race, education level, and health literacy [128]. Finally, further research should focus on determining which functionalities are mostly responsible for the effects on the outcomes.

The use of patient-centered digital health records in chronic conditions is potentially associated with beneficial effects on several patient-reported outcomes and recommended care services in a considerable number of studied digital health records. The rates of the effects were approximately similar for different patient groups. Feasibility and acceptability were high. Our findings support further implementation of patient-centered digital health records in clinical practice. Yet, higher-quality research is needed to identify effects per disease category and per health outcome and to learn which patients might benefit from specific functionalities.