This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Advance care planning allows patients to articulate their future care preferences should they no longer be able to make decisions on their own. Early advance care planning in outpatient settings provides benefits such as less aggressive care and fewer hospitalizations, yet it is underutilized due to barriers such as provider time constraints and communication complexity. Novel methods, such as patient portals, provide a unique opportunity to conduct advance care planning previsit planning for outpatient care. This follow-up to our pilot study aimed to conduct pragmatic testing of a novel electronic health record-tethered framework and its effects on advance care planning delivery in a real-world primary care setting.
Our intervention tested a previsit advance care planning workflow centered around a framework sent via secure electronic health record-linked patient portal in a real-world clinical setting. The primary objective of this study was to determine its impact on frequency and quality of advance care planning documentation.
We conducted a pragmatic trial including 2 sister clinical sites, one site implementing the intervention and the other continuing standard care. A total of 419 patients aged between 50 and 93 years with active portal accounts received intervention (n=200) or standard care (n=219). Chart review analyzed the presence of advance care planning and its quality and was graded with previously established scoring criteria based on advance care planning best practice guidelines from multiple nations.
A total of 19.5% (39/200) of patients who received previsit planning responded to the framework. We found that the intervention site had statistically significant improvement in new advance care planning documentation rates (
Advance care planning previsit planning using a secure electronic health record-supported patient portal framework yielded improvement in the presence of advance care planning documentation, with highest improvement in active patient portal users and patients aged between 50 and 60 years. Targeted previsit patient portal advance care planning delivery in these populations can potentially improve the quality of care in these populations.
Advance care planning (ACP) is the formal process of outlining a patient’s future care preferences should they lose the ability to make informed decisions for themselves [
The Institute of Medicine recommends conducting ACP early in a patient’s chronic disease diagnosis, with periodic reassessment every several years or with change in prognosis (such as new diagnosis, hospitalization, or worsening of chronic disease) [
EHR-linked patient portals, first described in the 1990s and now ubiquitous due to EHR Meaningful Use guidelines [
In an earlier study, our research team developed and pilot-tested a concise EHR patient portal–linked, electronic ACP communication framework in a small randomized controlled trial [
The aim of this study was to determine the impact of previsit ACP planning using a secure EHR-linked framework upon ACP documentation when incorporated into a real-world primary care environment.
This study was approved by the Ohio State University Institutional Review Board.
Patients 50 years or older, presenting for a preventive health or chronic disease follow-up visit, with an active MyChart account, at a participating clinical site were included in the study. Patients did not need prior MyChart experience. There were 2 clinical sites participating in the study. Sites were selected based on their demographic, size, and provider similarity, as well as their uniform clinical practices with respect to ACP delivery. Each clinical site used the same ACP practices before the study period, which included an institutional packet of information on ACP, state-issued documents about Advance Directives (ADs), and encouragement to discuss any ACP questions with their provider. The usual care site maintained these practices throughout the duration of the study. The intervention site incorporated an open-ended ACP framework (containing 4 questions), sent via a patient’s EHR-tethered patient portal, into a clinical practice algorithm. Physicians, nurses, and other clinical staff at the intervention clinic were collectively involved in developing this ACP previsit planning algorithm that was rolled out practice wide over a 3 month period (
Participants did not know their intervention was the intervention of interest. Each clinical site agreed to participate in an ACP process study. However, providers and patient participants did not receive labels about whether they were receiving intervention or usual care. Practice workflow was implemented without labels.
Data security was paramount in this study. We used clinical staff (Institutional Review Board approved) routinely interacting with the patient record and completing previsit planning for clinical care to administer the intervention. Our research team was embedded in the clinical site. Data were housed within the secure institutional firewall and only accessed within the clinical site. Only de-identified datasets were shared with the statistical team for analysis using a secure, institutional drive. The delivery system developed by the practice providers had built-in safeguards for addressing clinical emergencies, such as patients responding to the secure message with medical complaints, by having a nurse and physician on call for urgent messages.
Charts were reviewed both 1 week and 4 weeks post appointment. Charts were reviewed by a member of the team who had received training and quality control checks on ACP chart review protocol. Chart review findings were spot checked by a second member of the team (one every 20 records) to ensure accuracy. The protocol outlined that training and education interventions would be used to respond to discrepancies in chart review rates and quality assessments. However, interventions were not needed because spot checking did not yield discrepancies. The participant’s demographics, presence of ACP (including before and after the visit), quality of ACP if present, and number of MyChart messages sent in the last year were recorded. The intervention charts were reviewed to see if the patient had read the intervention on their portal and responded to any of the questions.
Intervention workflow that was implemented at the study practice during the 3-month trial period. ACP: advance care planning; PCP: primary care physician.
Demographic information from both the control and intervention clinic.
Demographic information | Control (n=219) | Intervention (n=200) |
Median age, years | 61 | 61 |
Age range, years | 50-93 | 50-91 |
Male, n (%) | 76 (35) | 101 (50.5) |
Female, n (%) | 143 (65) | 99 (49.5) |
Number of chronic diseasesa | 4 | 4 |
Number of medicationsa | 7 | 6 |
aMedian number per patient at each practice site.
Quality of ACP was measured using a 20-point scoring criteria entitled “Criteria for Scoring Quality of ACP Documentation” (
Our study analyzed the documentation rates and quality of ACP across both study sites, especially focusing on new ACP documentation appearing in the EHR. To assess whether or not the increase in new documentation was significant between the 2 sites, a Fisher exact test was used. To analyze quality, a Mann-Whitney test was used to test the significance in new ACP quality between the 2 sites. The data were also analyzed by age and portal usage. Participants were separated into age groups by decade and portal usage was defined as either high or low, with high usage being more than 10 portal messages in 1 year.
Of the 200 patients who were sent the intervention, 156 read the message (78.0% read rate) on their portal and 39 responded (19.5% response rate) to at least one question in the framework (see
Documented advance care planning (ACP) in electronic health record. Documentation rates represent the percentage of charts that had any form of ACP, and quality is rated by the 20-point scoring criteria.
Patient characteristic | Control | Intervention | ||
219 | 200 | |||
Preintervention, n (%) | 129 (58.9) | 74 (37.0) | ||
Postintervention, n (%) | 130 (59.3) | 94 (47.0) | ||
Rate percentage increase, n (%) | 0.7 | 27.0 | ||
130 | 94 | |||
Quality rating postintervention, mean | 3.26 | 4.09 | ||
109 | 96 | |||
Preintervention, n (%) | 54 (49.5) | 27 (28) | ||
Postintervention, n (%) | 55 (50.4) | 37 (39) | ||
Rate percentage increase, n (%) | 1.8 | 37 | ||
55 | 37 | |||
Quality rating postintervention, mean | 2.81 | 3.75 | ||
82 | 67 | |||
Preintervention, n (%) | 51 (62) | 27 (40) | ||
Postintervention, n (%) | 51 (62) | 36 (54) | ||
Rate percentage increase, n (%) | 0 | 33 | ||
51 | 36 | |||
Quality rating postintervention, mean | 3.25 | 4.19 |
aACP: advance care planning.
Our intervention did not appear to affect the percentage of patients who had a scanned document in their EHR; both before and after the intervention, approximately 14% (28/200 and 7/47) of patients had a scanned directive at that practice. One patient brought in an Advance Directive to be scanned after responding to our framework.
ACP documentation in the EHR increased by 27.0% (74/200 to 94/200) during the study period at our intervention site, compared with a 0.7% (129/219 to 130/219) increase at our control site (
Patients aged between 50 and 60 years saw the greatest increase in ACP completion rates. At our intervention site, documentation rose 37% (27/96 to 37/96) as compared with 1.8% (54/109 to 55/109) at the control site. Comparatively, the 61-70 age group saw a 31% (29/76 to 38/76) increase in documentation rates, and the 71-80 age group saw a 6% (17/26 to 18/26) increase at the intervention site. Our control site had a 0% increase in each of those 2 age groups, but higher baseline rates of ACP completion before the study period (64% and 81%). In the intervention arm, there was only 1 patient between 80 and 89 years, and 1 patient over 90 years, so there were insufficient data to analyze this group. Individuals in the 50-60 age group, however, had slightly lower ACP quality as compared with the study population as a whole.
Those who sent more than 10 MyChart messages in 1 year were defined as “High Portal Users” and comprised approximately a third of the study group at each site. Documentation rose by over 33% (27/67 to 36/67) at the intervention site for this group. Comparatively, low portal users (10 messages or less) at our intervention site saw a 23.4% (47/133 to 58/133) increase in documentation rates.
In this study, we found that patients exposed to our framework were significantly more likely to have ACP documentation in the EHR and the quality of that documentation was better. This intervention benefits both the patient and the provider by providing another way for patients to think through the difficult decisions of how they envision their future care before their office visits. For patients who already used the patient portal, adding the framework would be a seamless integration into their usual part of their care. This tool was used most frequently by patients in the 50-60 age group and already active on MyChart. Targeting patients who are high users or in this age demographic to receive this tool can be a strategy for providing high-yield individualized previsit planning for ACP using patient portals. This intervention did not capture many different demographics, including all nonportal users, so developing other strategies to improve ACP documentation against cultural, technological, and demographic barriers must be used to ensure that there are improved outcomes for all and to continue to address existing health disparities in ACP documentation [
In terms of workflow, the framework requires a member of the care team to send out the MyChart message 3-5 days before the appointment. If a patient responded, answers were appropriately documented and sent to the patient’s provider. In our study, the messages were sent out by a clinic nurse who could also answer any follow-up questions the patient had and then route the message to the appropriate provider. With increased team-based, patient-centered medical home patient outreach before appointments in primary care settings, as well as ubiquitous use of patient portals for practices to adhere to meaningful use guidelines, these interventions can be disseminated to a wide array of primary care practices. Higher rates of ACP documentation resulted, while reducing time needed to have a complete ACP discussion with the patient during the office visit, as existing answers have already been recorded and the patient had preparatory time to articulate their wishes.
Previous studies have shown the benefits of previsit planning; if the provider has documentation of some of the patient’s future care preferences beforehand, there can be a more productive discussion with the patient during their appointment [
The study was not designed to elicit qualitative feedback from patients and providers to promote its pragmatic implementation. However, the participating site liked the delivery system enough to implement it as a permanent intervention. Furthermore, the participating site shared the intervention, which has now been disseminated to the wider net of associated primary care sites at the institution.
Baseline rates of ACP documentation at each site were different, as noted in the results section, with the usual care site having higher rates of completed ACP documentation at baseline. However, preintervention chart review at both sites allowed assessment of typical documentation rates, to determine the change in documentation rates in intervention versus control before and after the intervention period. Additionally, patients had to have an activated MyChart account to be included, which excluded a portion of the clinic population.
Incorporating the patient portal into ACP delivery is a promising way to increase completion rates and efficiently facilitate the conversation between the provider and the patient about their future wishes. This strategy may be more effective in patients familiar with patient portal use, who regularly use patient portal communication to access clinical care.
Advance Care Planning Quality Grading. Criteria for scoring quality of ACP documentation in the EHR.
Framework Questionnaire.
advance care planning
electronic health record
The project described was supported in part by the OSU College of Medicine Roessler Research Scholarship (MF).
SB-B, MF, LM, LP, AC, and SK were involved in the concept and design; SB-B, MF, LM, and AC were involved in the acquisition of data; SB-B, MF, LM, and LP were involved in the analysis and interpretation of data; SB-B, MF, LM, and SK were involved in the drafting of the manuscript; SB-B, MF, LM, LP, AC, and SK were involved in the critical revision of the manuscript for important intellectual content; MF and LP were involved in the statistical analysis; and SB-B was involved in supervision.
SB-B was involved in developing the tested framework that is now available for nonproprietary use. She was not involved in administering the intervention or completing the primary chart review.