Patient-generated health data: Impact on promoting patient-centered point of care tobacco treatment in patients with cancer

Abstract

Introduction: Guideline-based tobacco treatment is infrequently offered. Electronic health record-enabled patient-generated health data (PGHD) has the potential to increase patient treatment engagement and satisfaction.

Methods: We evaluated outcomes of a strategy to enable PGHD in a medical oncology clinic from July 1, 2021 to December 31, 2022. Among 12,777 patients, 82.1% received a tobacco screener about use and interest in treatment as part of eCheck-in via the patient portal.

Results: We attained a broad reach (82.1%) and moderate response rate (30.9%) for this low-burden PGHD strategy. Patients reporting current smoking (n = 240) expressed interest in smoking cessation medication (47.9%) and counseling (35.8%). As a result of patient requests via PGHD, most tobacco treatment requests by patients were addressed by their providers (40.6-80.3%). Among patients with active smoking, those who received/answered the screener (n = 309 ) were more likely to receive tobacco treatment compared with usual care patients who did not have the patient portal (n = 323) (OR = 2.72, 95% CI = 1.93-3.82, P < 0.0001) using propensity scores to adjust for the effect of age, sex, race, insurance, and comorbidity. Patients who received yet ignored the screener (n = 1024) compared with usual care were also more likely to receive tobacco treatment, but to a lesser extent (OR = 2.20, 95% CI = 1.68-2.86, P < 0.0001). We mapped observed and potential benefits to the Translational Science Benefits Model (TSBM).

Discussion: PGHD via patient portal appears to be a feasible, acceptable, scalable, and cost-effective approach to promote patient-centered care and tobacco treatment in cancer patients. Importantly, the PGHD approach serves as a real world example of cancer prevention leveraging the TSBM.

Keywords: Patient-generated health data; cancer prevention; health informatics; tobacco treatment; translation.

Figure 1.

Patient-generated health data: pre-appointment tobacco screener during eCheck-in via the patient portal (MyChart) in electronic health record. The tobacco screener is delivered to outpatient return oncology visits every 90 days to ask patients about tobacco use, offer education via handout, and assess interest in smoking cessation treatments. Then, the clinical staff team assists in prescribing treatments and ordering counseling referral. The automatic system arranges for reassessment at next return visit 90 days later.

Figure 2.

Patient-generated health data: Clinical workflows for patients and providers.

Figure 3.

Patient-generated health data: Patients share treatment interest and receive tobacco treatment. response rate: Of the 10,496 screeners sent, 7,253 did not respond (69.1%) and 3,243 responded (30.9%). Tobacco use status: 1093 reported former smoking, 1906 never smoking, 156 current daily smoking, and 88 current someday smoking. Of 244 who screened positive for tobacco use (some day or daily smoking status), 4 did not answer treatment interest question. Interest in treatment: 240 answered treatment interest question. Of the 115 patients interested in medication, 19 were already receiving medication and 96 are newly interested. Of the 86 interested in counseling, 10 were already receiving counseling and 76 are newly interested.

Figure 4.

Tobacco treatment in three patient groups: Patients responding to screener, patients receiving/not responding to screener, and patients without the patient portal. (1) Patients responding to screener vs. Patients not having the patient portal are associated with more treatment received in multivariate logistic regression (OR, 2.72, 95%; CI, (1.94–3.82); P < 0.0001), adjusted for age, sex, race, insurance status, and comorbidity. (2) Patients receiving by not responding to screener vs. Patients not having the patient portal are associated with more treatment received in multivariate logistic regression (OR, 2.20, 95%; CI, (1.68–2.86); P < 0.0001), adjusted for age, sex, race, insurance status, and comorbidity.(3) Patients responding to screener vs. Patients receiving but not responding to screener are associated with more treatment received in multivariate logistic regression (OR, 1.16, 95%; CI, 0.86–1.58); P = 0.33), adjusted for age, sex, race, insurance status, and comorbidity. (4) N = 309 (group A, patients who smoked and responded to screener), N = 1024 (group B, patients who smoked and did not respond to screener), N = 323 (group C, patients who did not have the portal).

Figure 5.

Using the Translational Science Benefits Model to document patient-centered tobacco treatment.