Table 2.
Summary of the findings on health care efficiency.
| First author, year | Country | Health information technology | Study design | Sample size | Main findings |
| Ancker, 2019 [34] | United States | Blood glucose flow sheet (EpicCare and Weill Cornell Connect portal) | Observational, hypothesis testing | 53 patients | Uploaders had more clinical visits and portal logins before initial data upload. |
| Bidmead, 2016 [48] | England (United Kingdom) | Patients Know Best (PKB) | Descriptive, qualitative | 56 patients | The portal enabled clinicians to manage stable patients, facilitating clinical and cost-effective use of specialist nurses, and improved two-way communication and more optimal use of outpatient appointments and consultant time. It also facilitated a single rationalized pathway for stable patients, enabling access to information and proactive support. |
| Fiks, 2015 [60] | United States | MyAsthma (clinical interface in MyChart) | RCTa | 60 families of children | The intervention group had a marginally significant reduction in the proportion of parents missing at least 1 day of work (reduction of 47%, P=.07). Families in the intervention group reported fewer EDb visits and hospitalizations for asthma over 6 months than the control group (3 vs 9 and 0 vs 2, respectively). Only two intervention families reported at least one ED visit (vs six control families), and no intervention families reported hospitalizations. Children in the intervention group had fewer visits with asthma specialists or primary care. Results were similar on stratifying by asthma severity. |
| Fiks, 2016 [32] | United States | MyAsthma | Descriptive, mixed methods | 237 families | Portal users with uncontrolled asthma had significantly more primary care asthma visits after using the portal than the year earlier (increases of 16%). |
| Foster, 2019 [43] | United States | Epic MyChart | Observational, hypothesis testing | 208,635 tests | ED visits: 80.56% (n=20,430) of patients had a single ED visit with laboratory testing, 16.04% (n=4069) had two or three ED visits, 3.16% (n=802) had four to 10 ED visits, and only 0.24% (n=60) had more than 10 ED visits. Activation rates were lower for those with only a single ED visit (7312/20,430, 35.79%) compared with either those with two to three ED visits (1770/4069, 43.50%; P<.001) or four or more ED visits (368/862, 42.7%; P<.001). |
| Griffin, 2016 [44] | United States | My UNC Chart | Observational, hypothesis testing | 2975 patients | The odds of being readmitted within 30 days for active users was 66% higher than that for nonusers, holding all other variables constant in the model. There was no significant difference in 30-day readmission between nonusers and light users. |
| Jahn, 2018 [49] | United States | My HealtheVet | Descriptive, qualitative | 29 participants | Secure messaging tasks were inefficient as related to clinical document sharing (it took almost 5 minutes for providers to only attach and send a clinical document). |
| Kipping, 2016 [35] | Canada | Ontario Shores HealthCheck Patient Portal | Observational, hypothesis testing | 91 patients | Fewer missed appointments and a reduced number of requests for information in the year following portal implementation. The odds of a portal user attending an appointment were 67% (CI 56%-79%) greater than for nonusers over the follow-up period. Compared with 2014, in 2015, there was an 86% and 57% decrease in requests for information among users and nonusers, respectively (61% overall). |
| North, 2014 [38] | United States | Mayo Clinic Health System | Observational, hypothesis testing | 2357 primary care patients | Primary care patients who sent at least one secure message or e-visit had a mean of 2.43 (SD 2.3) annual face-to-face visits before the first message and 2.47 (SD 2.8) after, with a nonsignificant difference (P=.45). After adjustment for a first message surge in visits, no significant visit frequency differences were observed (mean, 2.35 annual visits per patient both before and after the first message; P=.93). Subgroup analysis showed no significant change in visit frequency for patients with higher message utilization or for those who had used the messaging feature longer. |
| Plate, 2019 [39] | United States | MyChart; Epic Systems Corporation | Observational, hypothesis testing | 6426 patients | Active MyChart status was not associated with 90-day ED return (P=.78) or readmission (P=.51) based on univariable analysis. Similarly, during multivariable analysis controlling for age, gender, BMI, and ASAc category, active MyChart utilization was not significantly associated with 90-day ED visits (ORd 1.019, 95% CI 0.843-1.231; P=.85) or readmissions (OR 0.966, 95% CI 0.747-1.249; P=.79). Patients who sent secure messages within 90 days from surgery (2200 patients, 48% of active users) were not less likely to present to the ED (P=.63) or be readmitted (P=.59) within 90 days. For patients who sent two or more messages (1354 patients), provider or staff response rate <75% was significantly associated with 90-day readmission (P=.004) with greater 90-day ED visits that neared statistical significance (P=.07). |
| Quanbeck, 2018 [54] | United States | Seva | Interventional, other than RCT | 268 patients | Significant reduction in hospitalizations and a trend toward fewer ERe visits. Increase in HIV screening rates. Change in the rates of HIV risk behaviors (eg, condom use) and receiving other addiction treatments appeared to be nonsignificant. |
| Riippa, 2015 [55] | Finland | Patient portal by The Finnish Medical Society, Duodecim | Interventional, other than RCT | 876 patients | The effect on the cost of care was ambiguous; costs decreased by an average of €91 in the unadjusted model, but increased by €48 in the adjusted model. Due to the controversial result, the unadjusted analysis showed an 89% probability of cost-effectiveness with no willingness to pay for increased patient activation, whereas in the adjusted sample, the probability of the portal being more cost-effective than care as usual exceeded 50% at a willingness to pay €700 per clinically significant increase in the patient activation score. For doctor visits, portal access (n=80): 3.8 (SD 3.3) and control (n=57): 3.0 (SD 3.1) (t=1.4; P=.18). For nurse visits, portal access (n=80): 3.5 (SD 2.6) and control (n=57): 4.1 (SD 2.5) (t=−1.3; P=.18). |
| Tsai, 2019 [28] | United States | Epic’s personal health record system | Descriptive, quantitative | 109,200 patients | Active users had more outpatient and inpatient visits and fewer ER visits. Patients without a portal account had on average fewer outpatient visits per month (0.31 vs 0.89, P<.001) and fewer inpatient visits per month (0.007 vs 0.059, P<.001), but had more ER visits per month than patients who were active with the portal (0.047 vs 0.014, P<.001). The difference between no-show appointments was not significant. |
| Wallace, 2016 [58] | United States | MyChart by Epic health record system | Interventional, other than RCT | 36,549 patients | The number of visits for 12 months was strongly associated with an increased likelihood of MyChart activation and with more frequent MyChart logins. |
| Zhong, 2018 [42] | United States | MyUFHealth (also known as MyChart by Epic) | Observational, hypothesis testing | 15,659 nonusers and 5494 users | At the time of adoption, the quarterly PCPf office visit RRg of users to nonusers was 1.33 (95% CI 1.27-1.39; P<.001). The RRs were between 0.94 and 0.99 up to four quarters after portal adoption (P=.75, .10, .13, and .09, respectively), and it was significantly less than 1 at the seventh (RR 0.82, 95% CI 0.73-0.91; P<.001) and eighth (RR 0.80, 95% CI 0.70-0.90; P<.001) quarters post adoption. The no-show rate proxies in the user group were significantly lower than in the nonuser group. RRs were between 0.60 and 0.83 for eight out of 11 quarters, and for the remaining three quarters, differences were not significant (P=.65, .29, and .44, respectively). Differences in cancellation rate proxies were not significant (P>.05). Overall, appointment adherence improved after portal adoption. |
aRCT: randomized controlled trial.
bED: emergency department.
cASA: American Society of Anesthesiology.
dOR: odds ratio.
eER: emergency room.
fPCP: primary care physician.
gRR: rate ratio.