Update 2022-2023

Greetings and welcome to the second update from The Permanente Medical Group’s Delivery Science and Applied Research (DARE) program!

TPMG’s DARE program provides infrastructure, connections, and analytic support to clinician-researchers who are answering questions that will change care within Kaiser Permanente.

Started in 2018 under the leadership of TPMG’s Chief Executive Officer and Associate Executive Directors, it is a collaboration between clinician investigators and Division of Research scientific experts, to inform large numbers of rapid-cycle, evidence-based innovations.

The DARE program includes:

  • A DARE support team to consult on ideas to facilitate successful clinician-led research;
  • A one-stop idea/funding support submission process to provide consistent, rapid, centralized project idea review and support for three DARE project funding mechanisms;
  • The TPMG Physician Researcher Program, which provides high-level support to selected clinician investigators to conduct clinically relevant research;
  • Specialty-specific research networks that build communities of evidence-driven clinicians. 
  • Centrally available tools, training, connections, and collaborations to speed the process from research results to implementation;
  • Expanded financial support for disseminating results in publications and national meetings.​

In the following pages, we invite you to explore these resources and the “DARE Dozen”: twelve sample DARE ​supported projects drawn from approximately 120 investigations underway across different specialties in 2021–2023 alone. These projects are poised to change care or have already done so; we hope you enjoy reading about these and invite you to learn more about others included here and at kp.org/dare​.

We look forward, in 2024-2025, to further supporting TPMG’s ongoing strategic areas of focus. It has been an active and exciting last two ears — the incredible talent and productivity of TPMG’s clinician-investigators and their dedication to evaluating and improving care is inspiring. We look forward to supporting your ideas, and always welcome new recommendations to improve.

Warm regards on behalf of the entire ​DARE team,

Douglas Corley, MD, PhD
Director, Delivery Science and Applied Research, TPMG

Tracy Lieu, MD, MPH
Director, Kaiser Permanente Division of Research

Yi Fen Irene Chen, MD and Smita Rouillard, MD
Associate Executive Directors, TPMG


Featured Projects - "DARE Dozen"​

AuthorsAssiamira Ferrara, MD, PhD, Mara Greenberg, MD
ChallengesPrenatal care visits are extremely common and direct important screening tests for pregnancy-related hypertension and diabetes. Telehealth may be a useful approach for these routine visits, but its impact on birth outcomes is unknown.
Existing EvidencePregnant patients and their providers perceive the use of telemedicine as a positive experience, though its effectiveness for prenatal care outcomes is unknown, including its potential role for decreasing differences in accessing care by location, demographics, and socioeconomic status.
Target PopulationIndividuals pregnant during and before the COVID-19 pandemic
Intervention or ExposureVisit type: office only, partial telehealth, full telehealth (pre-pandemic vs. increased telehealth during pandemic); and COVID infection from 30 days before conception to 7 days after delivery
Outcomes/Key FindingsA hybrid model of prenatal care using a combination of in-office and telemedicine visits is effective and could be used beyond the pandemic period. A higher (25%) proportion of prenatal visits from telemedicine intra-pandemic vs. 14% pre-pandemic provided similar comparable detection rates of gestational diabetes, preeclampsia, preterm birth and NICU admission rates; no differences were seen between people of different race, ethnicity, or measures of neighborhood deprivation.
Pregnant individuals with SARS-CoV-2 infection had higher risk of severe maternal morbidity (HR 2.45, 95% CI:1.91, 3.14), preterm birth (<37 weeks; 2.07 [1.75, 2.46])
and venous thromboembolism (3.08 [1.09, 8.74]).
Resulting Action/ChangeThese data support a next-step KPNC pilot of a nationally designed hybrid prenatal care model for low risk pregnancies.
Additional RecommendationsEvaluating other pandemic-associated prenatal care changes, including decreased early-pregnancy screening for gestational diabetes and change from multimodal aneuploidy screening to cell-free DNA screening to inform their inform continuation of these novel care patterns.
Implementation ToolsTelemedicine contact, device, and screening protocols for prenatal care.
Implementation and
Follow-up Measures
Proportions of pregnancies using telemedicine or hybrid approaches; completion of recommended screenings; detection rates of gestational diabetes, pre-eclampsia,
and diagnosed depression; differences by demographics and medical center.
AuthorsLisa J. Herrinton, PhD; Keras Lo, PharmD; Mubarika Alavi, MS; Stacey E. Alexeeff, PhD; Kerri M. Butler, PharmD; Carter Chang, MD; Christopher C. Chang, PharmD; Virginia L. Chu, PharmD; Ashok Krishnaswami, MD, MAS; Lynn H. Deguzman, PharmD; Stephanie Prausnitz, MS; Maisha Draves, MD, MPH; Michael D. Mason, MD
ChallengesMany older patients use large numbers of medications, which may increase the risk of adverse events without benefit. A proposed deprescribing intervention with pharmacists may help; however, its effectiveness is unknown, it is resource intensive, and it may create unanticipated adverse events from stopping medications.
Existing EvidenceDeprescribing has been implemented in numerous settings, including currently with Kaiser Permanente, but typically it involves only single drug classes. Using multiple single drug deprescribing efforts is inefficient, but little data exist on the potential beneficial or harmful impacts of “bundled” multi-drug deprescribing strategies.
Target PopulationKPNC members aged ≥76 years using ≥10 prescription drugs, excluding those with a history of transplant; dialysis; hospice care; or a cancer diagnosis, oncology visits, or cancer treatment within a year.
Intervention or ExposureA randomized trial of bundled multi-drug deprescribing using integrated pharmacist review of >20 drug classes within a single workflow, developed with relevant medical specialists, and using random patient assignment.
Outcomes/Key FindingsA large pragmatic clinical trial of pharmacist-directed multi-medication evaluation and patient counseling/deprescribingresulted in:
• no significant differences at 6-12 months in overall medication use (average within-person change, 0.4 medication, p=0.71)
• no beneficial reductions in geriatric syndrome conditions (difference-in-difference, 1.0, p=0.65)
• no increase in adverse effects from discontinuation. These results indicate no likely beneficial effect, beyond current pharmacy programs, for this type of intervention.
Resulting Action/ChangeInformed value-based care to not proceed with a planned large, complex intervention that would have required >10 pharmacist equivalent positions and complex coordination between pharmacy and adult & family medicine for full implementation.
Additional RecommendationsPragmatic trials prior to full implementation are feasible and have clear benefit in understanding effectiveness and to evaluate for potential unanticipated harms.
Implementation ToolsDrug protocols, medication discontinuation/continuation guidelines for disease-specific essential conditions, monitoring protocols/variables for adverse effects
Implementation and
Follow-up Measures
Drug protocols, medication discontinuation/continuation guidelines for disease-specific essential conditions, monitoring protocols/variables for adverse effects

AuthorsAdrian Hinman, MD, David Lee, M.D., Kamran Sadr, M.D., Kasey Cortese, M.D., Catherine Lee, Ph.D., Stephanie Prausnitz, M.S.., Andrew Avins, M.D., M.P.H.
ChallengesAdductor canal catheters are a common, invasive, and potentially beneficial method for improving post-operative pain control among patients undergoing total knee arthroplasty, but it is unknown if their use provides benefits beyond simpler pain management techniques, such as intra-operative peri-articular injection.
Existing EvidenceIn the last 7 years, the length of stay for total knee replacement patients in KPNC decreased from 3 days to <1 day, partially from improved pain control. Pilot KPNC data, and external trial data suggest adductor canal catheters may not substantially influence length of stay, opioid use, or pain scores beyond peri-articular injection.
They are also associated with complications (e.g. leaking, dislodging, bleeding, hematoma, prolonged neuropraxia), that can require additional outpatient & ED visits, and substantial resources.
Target PopulationAll patients undergoing a non-complex unilateral total knee replacement who are eligible to receive an adductor canal catheter
Intervention or ExposureUse of an adductor canal catheter as a supplement to other ERAS pain control protocols
Outcomes/Key FindingsPain control and postoperative use of opioid medications were comparable (non-inferior) between patients using vs. not using adductor canal catheters as a supplement to other pain management strategies after knee replacement surgery.
Complex and invasive multi-modal pain treatments using adductor canal catheters may not provide additional benefit for pain control, not impact length of stay, and increased costs/effort beyond simpler current methods for pain management
Resulting Action/ChangeThe orthopedic and pain control services are evaluating whether to modify the current common use of adductor canal catheters in patients undergoing knee replacement. This would be projected, for similar outcomes, to result in: comparable pain control; fewer catheter-related complications; decreased operating room staff, anesthesiologist, & room time for insertion; decreased nursing time for patient education; fewer post-operative visits for complications and removal; and directly reduced catheter costs of $1.5 million annually for KPNC.
Additional RecommendationsFor additional conditions, comparative evaluations of whether invasive multi-modality pain interventions may provide value beyond simpler interventions
Implementation ToolsData summaries regarding comparable outcomes for patient & physician education
Implementation and
Follow-up Measures
Adductor canal catheter use, variation by medical center, physician, and patient demographics; post-discharge healthcare utilization (e.g. ED, office, phone visits)

Manuscript Submitted
AuthorsJonathan E. Volk, MD MPH, Michael Silverberg, PhD MPH
ChallengesMany patients who are at risk for acquiring HIV would benefit from HIV preexposure prophylaxis (PrEP), but use remains limited and inequitable. Strategies to improve PrEP provision in primary care are urgently needed.
Existing EvidencePrEP uptake remains limited and inequitable, despite having a Grade A recommendation from USPSTF; in 2021 PrEP was prescribed to only 30% of those likely to benefit. Flagging the 2% of the general patient population with the highest HIV risk scores, our previously developed model prospectively identified 39% of new HIV diagnoses (46% among males), substantially outperforming CDC criteria for PrEP use, particularly among Black patients, a population with high HIV incidence and low PrEP uptake.
Target PopulationNon-HIV infected adult patients at KP San Francisco not previously prescribed HIV PrEP who have an estimated 3-year HIV incidence of ≥0.2% based on our EHR HIV prediction model.
Intervention or ExposurePrimary care providers at KP San Francisco, randomized to usual care or intervention (alerted via EHR-based staff messages prior to upcoming in-person or virtual visits and prompted to discuss HIV prevention and PrEP).
Outcomes/Key FindingsA low-intensity EHR-based intervention guided by an HIV prediction model:
• More than doubled PrEP prescribing/uptake among PCPs who provide care to both HIV-infected and uninfected patients (HR 2.59 (95% CI 1.30, 5.16)
• Did not change PrEP prescribing among providers who do not routinely provide care to HIV-infected patients (HR 0.89 (95% CI 0.59, 1.35).
Resulting Action/ChangeThe prediction score was integrated into the EHR and can be leveraged in the future, for patient outreach (e.g. outreach for e-visits, including the sexual health e-visits), to pharmacist co-management, and to enable providers to refer to PrEP with fewer barriers (e.g. best practice alert that emails patients links to PrEP e-visit). The risk score incorporated real-time patient data and facilitated provider notifications within routine workflows, informing future best practice alerts and other interventions to increase PrEP uptake.
Additional RecommendationsThese results support evaluating targeted applications of this risk tool and integrating with newly developed virtual tools such as e-visits for STI testing and PrEP and exploration of other methods for providers who do not routinely provide HIV care (and their patients).
Implementation ToolsHIV risk calculator integrated into the EHR.
Implementation and
Follow-up Measures
Initiation of PrEP care defined as PrEP discussions,
referrals, or prescription fills.

Manuscript Submitted
AuthorsAlan Go, MD, Amir Axelrod, MD, Andrew Ambrosy, MD; Jitesh Vasadia, MD
ChallengesIt is unknown whether novel, new, remote monitoring methods utilizing a comprehensive heart failure disease management program that includes telemonitoring and patient education might influence both health care utilization (such as emergency department care and hospitalization) and mortality
Existing EvidenceFour KPNC facilities (North Valley, South Sacramento, Santa Rosa and Napa-Solano) have used telemonitoring to follow and treat selected high-risk patients after a clinical encounter for worsening heart failure (HF). The benefit of telemonitoring from randomized trials is unclear, some reported potential clinical and resource utilization benefits, but most had small sample sizes (<900 patients) and were heterogenous in the type of telemonitoring.
Target PopulationEligible adults with heart failure at Roseville, Sacramento, South Sacramento, Santa Rosa, and Napa-Solano facilities between 2015-2019.
Intervention or ExposureUse of a natural experiment within KPNC, with some patients receiving remote home-based telemonitoring using the Residio LifeStream system. Patients with & without telemonitoring were matched on age, sex, race/ethnicity, service area, left-ventricular ejection fraction, HF duration, and a high-dimensional ropensity score.
Outcomes/Key FindingsAdults with HF who received telemonitoring had more frequent medication adjustments but comparable rates of hospitalizations and emergency department visits for worsening HF and all-cause mortality than controls who did not receive telemonitoring
Resulting Action/ChangeThese results do not definitively recommend widespread adoption of telemonitoring for patients with heart failure, given these findings and uncertain benefit in randomized trials.
Additional RecommendationsAn adequately powered randomized controlled trial may be necessary to establish confidence in benefits of telemonitoring for heart failure, overall or in targeted subgroups. These results reinforce the need for evaluating new devices and methods prior to widespread implementation.
Implementation ToolsNone
Implementation and
Follow-up Measures
Proportions of patient using telemonitoring, emergency department utilization, hospitalization rates, and mortality.

AuthorsJoseph Presti, Jr., MD, Stacey Alexeeff, PhD, Brandon Horton, MPH, Stephanie Prausnitz, MS, Andrew L. Avins, MD, MPH,
ChallengesCurrently only about 25% of men with mild PSA elevations have a prostate cancer that warrants treatment. Better risk assessment is needed.
Existing EvidenceNo prospective evaluation of a prostate cancer risk calculator has been performed in men with mild PSA elevations.
Target PopulationMen without prior biopsies who have mild PSA elevations (<10 ng/ml)
Intervention or ExposureUse of the validated KP Prostate Cancer Risk Calculator 2.0 in counselling for possible biopsy
Outcomes/Key FindingsCalculator usage led to a biopsy population enriched in high-grade cancer and fewer negative biopsies compared to historical controls and use for informed decision-making regarding prostate biopsy was accepted by patients and physicians.
• After hearing their estimated risks, 74% of men accepted biopsy
• Those that accepted biopsy had higher estimated risks for high-grade cancer and any cancer
• A threshold effect was seen at ~10% for high-grade risk and 30% for any cancer risk
Resulting Action/ChangeKP Prostate Cancer Risk Calculator can be considered for incorporation into HealthConnect for regular use in supporting clinicians and patients in shared decision making for prostate biopsies.
Additional RecommendationsThe KP Prostate Cancer Risk Calculator is currently being built into HealthConnect. We have initiated work with Garfield funding to further study our calculator in KPSC and to refine it using some novel serum markers that are not currently available in KPNC but are being used in KPSC.
Implementation ToolsThe risk calculator is available for next-step evaluation and use in different settings.
Implementation and
Follow-up Measures
Uptake of calculator, percentage accepting biopsy, biopsy outcomes.

Presented at the Western Section American Urological Association Meeting, October 3, 2023
AuthorsVarun Saxena, MD MAS; Liyan Liu, MSc; Nizar Mukhtar, MD; Sreepriya Balasubramanian, MD MPH; Brock MacDonald, MD; Suk Seo, MD; Joanna Ready, MD; Julie Schmittdiel, PhD, MA
ChallengesHepatocellular carcinoma (HCC) incidence has been the most rapidly increasing cancer for four decades. Little data exist for identifying high risk populations for recommended HCC screening, despite large at-risk KPNC patient populations with non-alcoholic fatty liver disease (NAFLD, estimated 200,000 members), treated hepatitis C (~13,000), and hepatitis B (HBV) (19,000 members).
Existing EvidenceFor chronic hepatitis B and treated hepatitis C, there have been a few non-US cohort studies of uncertain generalizability risk-stratifying patients HCC risk. The population of people with fatty liver disease is even larger.
Target PopulationCohorts of adults diagnosed between 2006-2018 with NAFLD, treated HCV after cure, or HBV were evaluated.
Intervention or ExposureFor each cohort, we evaluated the performance characteristics of externally derived risk scores for HCC (for HBV and treated hepatitis C); derived new KPNC risk scores for all cohorts; and then compared the performance of internal vs. external approaches (as applicable).
Outcomes/Key FindingsKP derived risk scores performed better than externally derived risk scores to stratify patients into low, medium, and high-risk groups, to accurately identify those likely to benefit / not benefit from every-6-month surveillance ultrasounds & serum AFP.  Among patients with NAFLD, the KPNC HCC risk score had an excellent ability (AUROC 0.88) to stratify patients into low-, medium-, and high-risk groups (Figure 1).  Among patients treated for hepatitis C (N=12,758), the KPNC HCC risk score(AUROC .94) was superior to external scores such as aMAP (0.74, without cirrhosis, 0.63 with cirrhosis) or the Toronto risk score (0.75 without cirrhosis, 0.65 with cirrhosis) (Figure 2).  Among patients with hepatitis B (N=30,287), the KPNC HCC risk score (AUROC 0.8) was superior to external scores (Figure 3) such as AGE-B (AUROC 0.82) and mPAGE-B (0.80).
Resulting Action/ChangeKPNC derived HCC risk scores will be implemented to use deploy a risk-based approach to HCC screening among patients with treated NAFLD, treated HCV, and HBV.
Additional RecommendationsWorking with KPNC IT to create real time HCC risk calculation within HealthConnect (or via toolbar), resultant HCC screening recommendations and population management
Implementation ToolsHCC risk stratification criteria that use electronic medical record variables.
Implementation and
Follow-up Measures
Use of risk stratification in patients with NAFLD, hepatitis C, and hepatitis B; cancer detection rates in patients by risk category

Manuscripts Underway
AuthorsVignesh A. Arasu, MD, PhD; Andrew L. Avins, MD, MPH; Laurel A. Habel, PhD; Jason D. Balkman, MD; Dorota J. Wisner, MD PhD; Dan Navarro, MD; Catherine Lee, PhD
ChallengesAt KPNC, over 100 radiologists read screening mammograms and demonstrate a 12-fold difference in cancer detection rates, including some below common guideline thresholds. Efficient time-to-exam reading times are challenged by volumes and staffing shortages, potentially inducing diagnostic delays.
Existing EvidenceRetrospectively, we demonstrated that high AI risk scores can identify 25-50% of breast cancers missed at screening. Whether it can achieve similar results prospectively and also help with real-time exam prioritization for diagnosing cancers same-day is unknown.
Target PopulationWomen undergoing mammographic screening (predominately aged 40–75 years)
Intervention or ExposureAI-based prioritization of mammography screening exam interpretation, including evaluation of 5 available commercial or open-source algorithms.
Outcomes/Key FindingsA retrospective evaluation demonstrated that:
• KPNC radiologists alone outperform AI algorithms alone (AUC 0.97 vs. AUC 0.83-0.91).
• Radiologist PLUS AI augmented cancer detection increased detection vs. radiologist alone, with AI identifying 25-50% of cancers not diagnosed by
radiologists.  A pilot prospective AI randomized controlled trial (RCT) in the Napa Solano service area found:
• Among 72 women diagnosed with breast cancer, AI correctly assigned 58 (80%) a high-risk score
• AI prioritization led to many more women with breast cancer potentially receiving same day results than without such prioritization.
Resulting Action/ChangeResults led to region-wide trial to evaluate improvements in cancer detection, informed a KP-IT computer vision AI use case for national strategy, and proof-of-concept technology and workflows, and informed data & technology infrastructure needs for at scale AI actions.
Additional RecommendationsThe results demonstrated the importance of prospectively evaluating technologies to inform adoption. Operational leaders can consider how results may inform TPMG AI strategy and development of proprietary workflows, as well as AI investments priorities.
Implementation ToolsEpic, Radiant
Implementation and
Follow-up Measures
Rate of rollout regionally, automated integration in workflow, cancer detection, time from mammogram to cancer diagnosis, cancer outcomes, diagnostic variability between radiologists after AI implementation

AuthorsKathryn K. Ridout, M.D., Ph.D., Chethana Eswarappa, M.D., Kelli Peterman, MS, Brooke Harris, Ph.D., Lyndsay Avalos, Dr.PH., Samuel J. Ridout, M.D., Ph.D.
ChallengesA large number of women with a history of depression take antidepressants, and many desire to stop medications during pregnancy. The impacts of continuing vs. stopping on the risk of post-partum depression (PPD) are unknown.
Existing EvidencePostpartum depression (PPD) impacts 1 in 7 women in the United States, is a large risk factor for maternal suicide, and impacts maternal-infant bonding and infant cognitive-emotional development. The American College of Gynecologists and American Psychiatric Association recommends women with a history of poor response to psychotherapy alone, moderate to severe PPD, or recurrent depressive disorders outside of pregnancy have initiation or continuation of an antidepressant during pregnancy.
Target PopulationWomen ≥18 with a live birth delivery between 2010-2019, an antidepressant fill in the 6 month prior to last menstrual period, and a depression diagnosis in the prior year
Intervention or ExposureContinuous antidepressant use during pregnancy.
Outcomes/Key FindingsWomen with a history of depression who discontinued (vs. continued use) were more likely to develop postpartum depression overall, (adjusted relative risk (aRR): 1.14; 95%CI 1.06-1.22 and 1.14) and had a higher risk of severe post-partum depression (PHQ-9≥20 within one year after delivery; aRR=1.33; 95% CI 1.09-1.62). Similarly, women who intermittently used their antidepressants during pregnancy had a higher risk of PPD overall (aRR=1.14; 95%CI 1.05-1.24). Regardless of antidepressant status, women with at least mild (PHQ-9≥5) depressive symptoms during the first depression screening of pregnancy had higher risk of PPD overall (aRR=1.49-1.75), and severe PPD (aRR=2.13-2.52). The number needed to treat to see the benefit of continued antidepressant use during pregnancy to prevent PPD is 15.
Resulting Action/ChangeAbility to inform patients and providers regarding antidepressant treatment in patients with pre-existing depression who are pregnant or planning pregnancy.
Additional RecommendationsDiscussions with patients based on their history of depression and depression severity regarding the relative risks and benefits could help improve patient outcomes.
Implementation ToolsRisk estimates, diagnostic criteria for identifying patients at risk.
Implementation and
Follow-up Measures
Percentage of patients continuing antidepressants during pregnancy, post-partum depression diagnosis.

Manuscript Submitted
AuthorsKian C. Banks, MD; Jennifer R. Dusendang, MPH; Julie A. Schmittdiel, PhD, MA; Diana S. Hsu, MD; Simon K. Ashiku, MD; Ashish R. Patel, MD; Lori C. Sakoda, MPH, PhD; Jeffrey B. Velotta, MD
ChallengesNational guidelines recommend large-scale lung cancer screening among at-risk patients and KPNC has instituted programs to fulfill this goal. However, the ideal time to surgery for early-stage lung cancer is unknown, including if currently recommended minimal diagnosis-to-surgery time recommendations accurately decrease the risk of stage progression.
Existing EvidenceThe time between lung cancer diagnosis by radiologic imaging and subsequent surgical resection varies widely. This variation is influenced by the need for additional work-up, patient preference, and operative scheduling. Few studies have evaluated if there is an optimal minimum time to surgery, beyond which there is an increased risk of recurrence or stage progression to inform patients, providers, and large-scale lung cancer screening programs.
Target PopulationStage I or II lung cancer patients aged 18 to 85 years undergoing elective surgery who had CT or PET imaging within six months prior to surgery
Intervention or ExposureTime from last CT or PET scan to surgery (TTS)
Outcomes/Key FindingsEarly-stage lung cancer patients with a time to surgery within 4 weeks experienced:
• lower rates of death (HR 1.18, 95% CI 1.00-1.39) and
• lower rates of cancer recurrence (HR 1.33, 95% CI 1.10-1.62).
This suggests the maximum time to surgical resection after lung cancer diagnosis to optimize outcomes may be much shorter than previously recommended, though
longer than current KPNC targets of within 2 weeks, which can be difficult to operationally achieve.
Resulting Action/ChangeThe results inform clinicians, patients, and large-scale lung cancer screening programs (such as those at KPNC) regarding evidence-based care systems &amp; targets for patients with early stage lung cancer. The findings will be shared with thoracic surgery; medical, surgical & radiation oncology; pulmonology; radiology and the regional screening program. The results may also modify current KP targets for surgery within two weeks, increasing flexibility for scheduling.
Additional RecommendationsThese results inform potential changes to national society guidelines on optimal time to treatment for lung cancer patients. This analytic approach is applicable to other screening programs early-stage cancers, to guide health-system level program development and planning.
Implementation ToolsData on time to surgery vs. outcomes, mechanisms for identifying at-risk patients
Implementation and
Follow-up Measures
Time to surgery, variation by medical center, cancer stage.

AuthorsLori C. Sakoda, PhD, MPH, Jie Zhang, PhD, Douglas A. Corley, MD, PhD, George Minowada, MD, Charles P. Quesenberry Jr, PhD, Sundeep M. Nayak, MD
ChallengesThe implications of implementing the 2021 U.S. Preventive Services Task Force(USPSTF) lung cancer screening guidelines on the screening-eligible population and for decreasing demographic differences in screening are largely unknown.
Existing EvidenceIn March 2021, the USPSTF revised their 2013 recommendations on lung cancer screening, expanding age eligibility to 55-80 years and smoking to ≥20 packyears,
in part to increase eligibility for more Black/African American and women patients. The impacts on detection, screening volumes, and disparities are unclear.
Target PopulationPopulation 1 - adults aged 50-80 years who ever smoked cigarettes, as of 06/30/2020; Population 2 -Adults diagnosed with incident primary lung cancer from
01/01/2014 to 12/31/2018.
Intervention or ExposureExpanded screening criteria by age and fewer pack years smoked.
Outcomes/Key FindingsImplementing new screening guidelines would:
• Almost 2x patients eligible for lung cancer screening from 11.4% (2013 criteria) to 20.4% (2021 criteria);
• Likely eliminate existing screening eligibility differences between Black and White patients diagnosed with lung cancer;
• Still screen only few of those ultimately diagnosed with lung cancer 21.9% (2013) vs. 28.9% (2021)
• Increase screening among people diagnosed with lung cancer who are Asian/Pacific Islander, Hispanic, and women, though differences compared with non-Hispanic White patients would persist.
Among ever smokers, changing eligibility criteria gains were similar across race categories (9.3%; 9.0%-9.6%); lower by Hispanic ethnicity (7.1%); and larger for men (9.9%) than women (7.9%). Among people diagnosed with a new lung cancer, new criteria eliminated prior screening eligibility differences between Black and White people (now 31.1% vs. 33.5%, respectively). Black people had the largest gains in eligibility (10.8%).
Resulting Action/ChangeThe new guidelines markedly expand the number of smokers eligible for screening. Healthcare systems will need to plan relevant resources to provide screening to the expanded pool of screening-eligible adults
Additional RecommendationsAdditional strategies for early detection of lung cancer are needed for at-risk patients, as most who develop lung cancer are not recommended for screening
Implementation ToolsNumber of people eligible for lung cancer screening in KPNC, for informing program resources & implementation
Implementation and
Follow-up Measures
Proportions of people identified as eligible for screening (smokers) from electronic records, proportions screened, differences by demographic groups.

Manuscript Submitted
AuthorsVeronica Shim MD, Audrey Karlea MS LCGC, Leslie Manace Brenman MD MPhil, CK Chang, Jamila Gul BS, Elizabeth Hoodfar MS CGC, Tracy D Chan RN, Poline C Engeman NP, Vanessa M. Sheldon, RN, Deirdre M Thorne-Hadfield NP, Patience Odele MD, Brooke Vuong MD, Jennifer McEnvoy MD, Dinesh Kotak MD, Laurel Habel PhD.
ChallengesGenetic testing for many patients with cancers is becoming standard of care. The current clinical pathway of universal genetic counselor referral for testing is challenged by both patient attrition in the referral process and feasibility. In addition, current NCCN guidelines for testing aimed at younger age groups may not detect many older breast cancer patients with genetic changes.
Existing Evidence“Mainstreaming” genetic testing within current clinical care pathways is a possible way to meet the increased need in genetic testing among breast cancer patients,
while personalizing their care with providers who are already familiar to them.
Target PopulationKPNC patients diagnosed with breast cancer
Intervention or ExposureBreast care coordinators for genetic testing order and pretest counseling at 4 centers (Central Valley, EB, San Rafael, S. Sac) vs. standard pathways for genetic counsel referral at 11 other centers. Within the pilot program, onlypatients with abnormal genetic profiles (including variants of unknown significance) were referred to genetics.
Outcomes/Key FindingsOffering genetic testing to all women under 65 with breast cancer through frontline breast cancer care coordinators vs. through genetic counselor referral:
• Doubled proportion completing genetic testing (from 31.7 to 61.6%)
• Decreased time to testing by 1/3 (from median 33 to 22 days)
• Identified 74% more people with pathologic variants just from expanding the age criteria (from 5.4% vs 3.1% of all women with cancer)
A substantial portion of patients had variants of uncertain significance. (46.3% at pilot vs 54.2% at non-pilot sites)
Resulting Action/ChangeThese findings recommend, among patients with breast cancer and potentially those with other cancers commonly associated with genetic syndromes:
• moving initial genetic testing to frontline providers, with referral to genetics targeted for those likely to benefit those with positive tests
• offering testing to all women with breast cancer up to at least age 65
Additional RecommendationsSteps for broader “mainstreaming” of genetic testing for patients with breast cancer include: 1) coordination with breast care coordinators or clinicians at additional centers 2) training coordinators with protocols from the pilot centers 3) incorporation of care flows in the newHealthConnect® genomic module
Implementation ToolsUpdating the HealthConnect® genetic module in 2023 by the national genomics team, to facilitate expansion of “mainstreaming” genetic testing
Implementation and
Follow-up Measures
Proportions of patients with breast cancer who complete genetic testing by age groups, proportions with positive tests, differences by demographics & medical center, and proportions who complete follow-up with genetic counseling for abnormal test results.

Manuscript Submitted

DARE Supported Projects 2021-2023​​

Delivery Science Grants Program
AnesthesiologyBrad Cohn, Vincent LiuEvaluation of an Electronic Health Record-Based Risk Stratification Algorithm for Perioperative Medicine Triage
Anesthesiology,  SurgeryEdward Yap, Mary ReedToward Reducing Surgical Complications After Urgent and Emergent Surgery
CardiologyDrew AmbrosyDefining the Role of ElectroNic HealTh Record-Based Alerts Worsening Heart Failure (IDENTIFY-WHF) Improves Care
CardiologyEd McNulty, Alan GoTranscatheter aortic valve replacement (TAVR) – Personalizing risk for QOL, complications, total mortality, and utilization
CardiologyAmir Axelrod, Andrew Ambrosy, Alan GoNo Differences in ED Visits, Hospitalizations, or Mortality Among Heart Failure Patients with Remote Telemonitoring
Cardiology,  Health Care Delivery and PolicyDustin Mark, Mary ReedValidation of an Algorithm for ED Patients with Possible Acute Coronary Syndromes After Transition to a High-Sensitivity Troponin Assay
Cardiology, Population CareAdam Rogers, Carlos IribarrenImproving quality by TARGETing LDL using a systems-based approach to evaluate our current performance (TARGET-LDL)
COVIDMary Reed, Dustin BallardPost-Acute Sequelae of SARS-Cov-2 Infection (PASC) in Adult KPNC Members
COVIDTracy LieuOutreach to Enhance COVID-19 Vaccination Among Hesitant Elderly
COVID,  Population CareGabriel Escobar, Laura Myers, Vincent LiuTools for Outpatient and Population Management of SARS-COV-2 Infections (TOPS2)
COVID, Ob/Gyn, Womens and Childrens HealthMara Greenberg, Assiamira FerraraA Prenatal Care Strategy that Included Telehealth had Similar Outcomes as In-Person Visits
Emergency MedicineDavid Vinson, Mary ReedReducing Variation in Hospitalization and  Care in ED Patients with A-fib: A Stepped Wedge Cluster Randomized Trial
Endocrinology,  Population CareLisa Gilliam, Richard GrantTesting a new population management model for hypoglycemia prevention in high-risk KPNC members
GastroenterologyT.R. Levin, Jeffrey Fox, Jeffrey LeeDe-implementation of Outdated Colonoscopy Surveillance Interval for Patients with Low-risk Adenomas (DESIRE)
GastroenterologyJeff Lee, Theodore LevinUnderstanding the Difference of Risk of Advanced Colorectal Neoplasia in Patients with Non-advanced Adenomas (3-4 vs 1-2)
GastroenterologyVarun Saxena, Julie SchmittdielNovel Risk Stratification Optimizes Hepatocellular Carcinoma Surveillance Protocols and Guidelines
Infectious Disease, Primary CareJonathan Volk, Michael Silverberg,Electronically Identifying Patients at Elevated Risk of HIV Increased Pharmacologic HIV Pre-exposure Prophylaxis Use: A Clinical Trial
Mental and Behavioral HealthHonor Hsin, Esti IturraldeDevelopment of a scalable and adaptable clinical workflow for machine learning (ML)-based suicide prevention in mental health care
Mental and Behavioral HealthKevin Li , Esti IturraldeEvaluation of electroconvulsive therapy, transcranial magnetic stimulation, and intravenous ketamine for treatment-resistant depression
Mental and Behavioral HealthKathryn Erickson-Ridout, Connie WeisnerEvaluation of the Achieving Depression and Anxiety Patient Centered Treatment (ADAPT PLUS) Program
Mental and Behavioral Health, Ob/GynKathryn Erickson-Ridout, Lyndsay AvalosEvaluation of the EmbrACE Intervention for Adults with a History of Adverse Childhood Experiences
NephrologySijie Zheng, Alan GoKidney Disease Population Care
NeurologyMai Nguyen-Huynh, Jeff KlingmanExtending Intravenous Tenecteplase Beyond 4.5-hour Window for Patients with Acute Ischemic Stroke
Ob/GynJoanna Stark, Lyndsay AvalosEvaluate “Centering Pregnancy”, a group prenatal care program, impact on health outcomes and health care utilization.
Ob/Gyn, Womens and Childrens HealthMara Greenberg, Monique HeddersonEvaluating patient engagement in "KP Connected Pregnancy Care" and its association with utilization and perinatal outcomes
OncologyRaymond Liu, Gabriel Escobar, Vincent LiuAutomatic risk stratification for hospital-acquired VTE to improve utilization of risk-appropriate pharmacological prophylaxis
OncologyCharles Meltzer, Lori SakodaEvaluating outcomes and efficiency of consolidated multidisciplinary care for head and neck cancer
OncologyBethan Powell, Larry KushiGenetic testing of women with hereditary breast and ovarian cancer syndrome
OncologyJed Katzel, Stephen Van Den EedenImplementation of Patient Reported Outcomes in Head and Neck Cancer
OncologyStephen Van Den Eeden,Physical Function in Pancreatic Cancer Patients
Oncology, Population Care, UrologyJoseph Presti, Andy AvinsImproving shared decision making for prostate cancer screening
Oncology, SurgeryVeronica Shim, Laurie HabelDeveloping a TPMG Pathway for Selective Oncotype Testing in Early Breast Cancer
Optometry/OphthalmologyDariusz Tarasewicz, Oleg SofryginImplementation of an evidence-based risk calculator for diabetic retinopathy screening and population management
OrthopedicsAdrian Hinman, Andy AvinsAdductor Canal Catheters not Superior to Conventional Therapy for Postoperative Pain Management After Total Knee Arthroplasty: A Trial
PediatricsMeghan Davignon, Lisa CroenEvaluation of More Than Words Program at KPNC
PediatricsJosephine Lau, Stacy SterlingIdentifying Best Practices of Early Phase Pediatric Eating Disorder Care
PediatricsLisa Chyi, Michael KuzniewiczUse of a modified Eat, Sleep, Console tool for management of Neonatal Abstinence Syndrome
PediatricsMichael Kuzniewicz, Dan BruggerPredictive Analytics for Newborn Bilirubin Management
PediatricsMustafa Bseikri, Elizabeth FelicianoPediatric Obstructive Sleep Apnea: Predictive Modeling to Streamline Care
Population CareLisa Gilliam, Andrew KarterComparative effectiveness of intermittently scanned continuous glucose monitoring vs. self-monitoring of blood glucose among insulin-treated type 2 diabetes patients
Population CareLisa Gilliam, Andrew KarterImplementation of evidence-based risk stratifiaction tools for diabetic retinopathy screening and population management
Population Care, Mental and Behavioral HealthLisa Fazzolari, Esti IturraldePopulation management for schizophrenia, schizoaffective disorder, and bipolar disorder utilizing trained psychiatric pharmacists
Primary CareMaisha Draves, Lisa HerrintonA Randomized, Multi-Drug De-Prescribing Intervention did not Change Medication Use or Geriatric-Syndrome Diagnoses.
Primary CareJoan Lo, Kendal HamannAssessment of fracture prevention quality measures
Primary CareSomalee Banerjee, Alyce AdamsImproving Outcomes and Care Experience among Dual Eligible Members: The Role of Health System Factors
Pulmonology,  Thoracic SurgeryEduardo Solbes, Laura MyersUse and outcomes of intrapleural fibrinolytic therapy for complicated parapneumonic pleural effusion and empyema
RadiologyVignesh Arasu, Laurel HabelProspective Mammography AI Evaluation Improved Cancer Detection and Time-to-Detection
SurgeryAmanda Graff-Baker, Marilyn KwanLymphatic Microsurgical Preventative Healing Approach to Prevent Lymphedema in Patients Undergoing Axillary Lymph Node Dissection for BrCA
UrologyJoseph Presti, Andrew AvinsNovel KP Prostate Cancer Risk Calculator Leads to Better Selection of Men Needing a Biopsy in Those with Mild PSA Elevations

Rapid Analytics Unit
AnesthesiologyEdward Yap, Lisa HerrintonAssessing variation in the incidence of myocardial injury after non-cardiac surgery (MINS) across medical centers and over time
AnesthesiologyJonathan Khersonsky, Cynthia CampbellAssociation of Nerve Blocks with Hospital Length of Stay for Surgical Abdominal Procedures
CardiologyKavin Desai, Lisa HerrintonAccuracy of Prenatal Detection of Clinically Significant Congenital Heart Disease Among Still- and Live-Births
CardiologyRichard Birnbaum, Andy AvinsFamilial hypercholesterolemia II - manuscript
CardiologyRichard Birnbaum, Andy AvinsFamilial Hypercholesterolemia Implementation
CardiologyUma Vadlakonda, Mary ReedImproved Outcomes with Home-Based Cardiac Rehabilitation-Experience from a Large Integrated Heathcare System
COVID, Mental and Behavioral HealthKathryn Ridout, Esti IturraldeMental Health Service Demand in the face of COVID-19
COVID, Physical and Rehabilitation MedicineJoshua Rittenberg, Julie SchmittdielAssociation Between Injected Corticosteroids and Risk for Influenza and COVID-19
Dermatology,  PharmacyVeena Vanchinathan, Andy AvinsOptimizing Acne Care: What is the Impact of Isotretinoin?
Dermatology, COVIDPatrick McCleskey. Lisa HerrintonEpidemiologic Analysis of Chilblains Eruptions at the time of COVID19
Emergency MedicineMamata Kene, Mary ReedImpact of Opioid Safety Initiative Education on Emergency Department Opioid Prescribing
Emergency MedicineDustin Mark, Mary ReedTrends in the use of computed tomographic cerebral angiography for emergency department patients with headache
Emergency Medicine, COVIDDale Cotton, Mary ReedCharacteristics of ED Patients Who Test Positive for SARS-CoV-2 in a Community Health Setting
GastroenterologyKrisna Chai, Joanna Ready, Andy AvinsAn Organized Hepatitis B Surveillance Program increases Patient Identification and Optimizes Management
GastroenterologyVaruna Saxena, Julie SchmittdielBenefits of Hepatitis C Virus Cure
GastroenterologyKrisna Chai, Joanna Ready, Andy AvinsImplementation success of integrated program on hepatitis B identification, surveillance, and treatment
Gastroenterology, COVIDFernando Velayos, Julie SchmittdielCOVID-19 Complications Are Not More Common Among Immunosuppressed Populations in KPNC
Gastroenterology, Oncology, Population CareSreepriya Balasubramanian, Julie SchmittdielEffect of Screening for Hepatocellular Carcinoma on cancer-related and all-cause mortality in patients with cirrhosis
Head and Neck, RadiologyKevin Wang, Andy AvinsImproving Diagnostic Efficiency of Adult Neck Masses
Infectious DiseaseFernando Velayos, Julie SchmittdielRisk of COVID-19 pandemic on immunosuppressed populations
Infectious Disease,  Hospital MedicineMeredith Silverman, Cynthia CampbellVancomycin area under the curve-guided dosing and monitoring in the outpatient setting
Mental and Behavioral HealthKathryn Ridout, Lyndsay AvalosContinuation of Antidepressant Use During Pregnancy Lowers Risk of Postpartum Depression
Mental and Behavioral HealthSamuel Ridout, Cynthia CampbellIntegration of mobile health (mHealth) based interventions in mental health
Mental and Behavioral Health,  Ob/GynCarey Watson, Cynthia CampbellThe impact of ACEs and resilience on Obstetric health outcomes
NeurologyJosiah Ambrose, Cynthia CampbellSurgical Epilepsy Center Referral Time Within an Integrated Health Care Network
Neurology,  Population CareJeff Klingman, Abby EatonStroke EXPRESS: Transition to TNKase
Ob/Gyn,  COVIDEve Zaritsky, Andy AvinsThe Impact of California Shelter-In-Place on Incidence of Gynecologic Emergencies in a Community-Based Population
Ob/Gyn,  Primary CareMehreen Khan, Mary ReedRemote Patient Monitoring Assessment
OncologyRobert Li, Lisa HerrintonChanges in utilization, staging, and treatment after implementation of regionalized care for upper gastrointestinal cancer
OncologyTom Urbania, Lisa HerrintonEvaluation of structured reporting of lung nodules detected on chest CT
Oncology,  COVIDAshok Pai, Gwendolyn Ho, Julie SchmittdielThe impact of the use of anticoagulants and antiplatelets on the risk of hospitalization and mortality in patients with COVID-19
Oncology,  SurgeryKourosh Kojouri, Lisa HerrintonPAN   3: Efficacy and Value
Oncology, Population CareRaymond Liu, Julie SchmittdielCancer outcomes with KPNC vs. national community averages: an overview study
Oncology, RadiologyLori Sakoda, Sundeep NayakImplementing Updated Lung Cancer Screening Guidelines Would Likely Decrease Demographic Disparities Found in Existing Practices
Optometry/OphthalmologyNaveen Chandra, Marilyn KwanDropless Cataract Surgery
Optometry/OphthalmologyI-Hui (Elaine) Wu, Mary ReedOutcomes of telemedicine in Ophthalmology
Optometry/OphthalmologyCatherine J. Choi, Lisa HerrintonNeurotrophic keratopathy: Incidence, presentation, utilization, and outcomes before adoption of corneal neurotization
OrthopedicsJohn Cox, Andy AvinsPerioperative VTE risk in total joint patients who had COVID-19
PediatricsJosephine Lau, Julie SchmittdielDelivery of Inpatient Pediatric Eating Disorder Care in a Large Integrated Health Care Delivery System
Pediatrics,  Hospital MedicineKelsey Childress, Marilyn KwanReducing Blood Culture Contamination and Subsequent Healthcare Reutilization in Pediatrics
Pediatrics, SurgeryAlbert Chong, Lisa HerrintonImplementation of Evidence-Based Surgery for Pediatric Hernia Repair
Physical and Rehabilitation MedicineMamie Air, Julie SchmittdielYearly corticosteroid burden of patients receiving spinal corticosteroid injections
Plastic SurgeryTodd Theman, Julie SchmittdielSafety and Efficacy of Corticosteroid Choice in Trigger Finger Injection
Primary CareKevin Wang, Janet Lai, Andy AvinsImproving Human Papilloma Virus Immunization Rates in Kaiser Permanente Northern California
Pulmonology, COVIDNareg Roubinian, Julie SchmittdielIncidence of outpatient thromboembolism in patients with SARS-CoV-2 infections
SurgeryReza Rahbari, Lisa HerrintonImplementation and Outcomes Following Regionalization of Adrenal Surgery
SurgerySimon Ashiku, Andy AvinsMinimally invasive esophagectomy
Surgery, COVIDGillian Kuehner , Mary ReedHow COVID Changed Our Practice: Telehealth in TPMG Surgical Cancer Care
Thoracic SurgeryJeffrey Velotta, Lisa HerrintonBody weight changes with gastrostomy or jejunostomy feeding tube, esophageal stent, or no intervention in esophageal cancer patients starting chemotherapy and chemoradiation
Thoracic SurgeryJeffrey Velotta, Julie SchmittdielIdentifying Surgical Timing to Optimize Outcomes for Early-Stage Lung Cancer Surgery
Vascular SurgeryNayan Sivamurthy, Julie SchmittdielAssessing the Impact of a Mobile Application Based Supervised Exercise Therapy (SET) Program for Peripheral Arterial Disease (PAD)

Physician Researcher Program
CardiologyMatthew Solomon, Alan GoOptimizing Management of Valvular Heart Disease
Emergency MedicineDavid Vinson, Mary ReedImproving management of ED patients with undifferentiated syncope; prospective validation of the Canadian Syncope Risk Score
Emergency MedicineDana Sax, Mary ReedImproving Risk Stratification of ED Patients with Acute HF using a Machine-learning Platform for Personalized, Accurate, Real-time Risk Prediction
GastroenterologyDan LiImproving health outcomes and reducing healthcare costs in individuals with Helicobacter pylori infection
GastroenterologyT.R. Levin, Jeffrey Lee, Vincent LiuPREDICT PCCC: PREDICTing Post Colonoscopy Colorectal Cancer
Infectious Disease,  COVIDJacek SkarbinskiImmunity and patterns of disease exposure for COVID-19 across Kaiser Permanente Northern California
Mental and Behavioral HealthMatthew Hirschtritt, Stacy SterlingA mixed-methods, longitudinal analysis of a novel telepsychiatry assessment, brief treatment, and referral clinic model
Mental and Behavioral HealthKathryn Erickson-Ridout, Constance Weisner, Esti IturraldeEvaluation of the ADAPT PLUS program as a model for care delivery to outpatients with depression and anxiety
NeurologyMai Nguyen-Huynh, Alan Go, Bruce OvbiageleExtended window for thrombectomy: Have we extended our resources appropriately and effectively?
Ob/Gyn, Population Care, Infectious Disease,Betty Suh-BurgmannHPV self-collection
RadiologyVignesh Arasu, Laurel HabelImproved selection of BRCA-negative high-risk women for breast MRI screening through validation of IBIS risk model variants
UrologyJoseph Presti, Andy AvinsRefining Shared Decision Making in Prostate Cancer Screening and Detection
Vascular SurgeryRobert Chang, Andy AvinsOptimizing Abdominal Aortic Aneurysm Surveillance and Intervention to Prevent Rupture
Targeted Analysis Program
Allergy/Immunology, COVIDLindsay Finkas, Carlos IribarrenRetrospective analysis of incidence and health outcomes of asthma patients with COVID-19 within a large multi-site healthcare system
Anesthesiology, COVID, Hospital MedicineLaura Myers, Kevin NgTrend in treatment location of patients with Covid-19 infection in an integrated health system
Emergency Medicine, Surgery, Thoracic SurgeryJeffery Velotta, Vinnie LiuComparion of outcomes of video assisted thoracic surgery (VATS) versus robot assisted thoracic surgery (RATS) for lobectomy
Endocrinology, Ob/Gyn, Womens and Childrens HealthMara Greenberg, Monique HeddersonRemote Glucose Monitoring in Pregnancies with Gestational Diabetes: Assessing Uptake and Outcomes
General Surgery, Genetics/Genomics, OncologyVeronica Shim, Laurel HabelStreamlined, Risk Based Multi-Gene Testing of Women with Breast Cancer in KPNC
Hospital MedicineMichael Yoo, Vinnie LiuEvaluation of physical function measure by the PF-5 in Kaiser Permanente Northern California
Mental and Behavioral HealthKevin Li, Esti IturraldeElectroconvulsive therapy, anesthetic choice, and real-world patient outcomes
Primary CareRajiv Misquitta, Issac ErgasBodyweight, cholesterol, and glycemic control among patients with CHD and/or type 2 diabetes: a diet and lifestyle intervention program
RadiologySundeep Nayak, Lori SakodaAssessing eligibility, participation, and initial outcomes in lung cancer screening
UrologyMark Gasparini, Stephen Van Den EedenSignificance of borderline high and high serum calcium levels in high risk kidney stone patients in a community practice