Penn LDI awarded 12 grants to teams led by LDI Senior and Associate Fellows engaged in work across a wide range of health services research areas. Read more below about the funded projects completed in 2020.


Jalpa Doshi
Jalpa Doshi, PhD

Predicting Prescription Drug Adherence and Treatment Gaps for Medicare Beneficiaries with Chronic Condition: A Comparison Between Traditional Models and Machine Learning Algorithms

Principal Investigator:  Jalpa Doshi, PhD | Perelman School of Medicine

Using 100% Medicare files from 2011—2014, we measured adherence to and treatment gaps in disease-modifying therapies (DMTs) in individuals with multiple sclerosis (MS) over a one-year and three-year follow-up. We found a downward trend in DMT adherence over time in both samples. Monthly adherence dropped significantly between December of the prior year to January of the following year (from 76% to 65% in the one-year follow-up sample and similar drops seen across all years in the three-year follow-up sample). Next, we used traditional regression models to identify predictors of DMT adherence and treatment gaps. In the one-year follow-up sample, several sociodemographic factors were associated with the odds of being adherent. Disability as the current or original Medicare entitlement reason was associated with lower odds of being adherent across all age categories. Men had higher odds than women and Black people had lower odds of being adherent compared to white people. In terms of clinical factors, patients with a higher RxHCC score and having an inpatient stay in the baseline year had lower odds of being adherent. Interestingly, patients with MS relapses in the baseline year had lower odds of being adherent in the follow-up year. In terms of Part D benefit design factors, patients in plans with the standard Part D benefit or plans that offered a basic alternative or actuarially equivalent to standard benefit design had lower odds of being adherent than patients in the more generous enhanced alternative Part D plans. Being in the high cost-sharing phase was associated with lower odds of being adherent than in the low cost-sharing phase.


Ari Friedman
Ari Friedman, MD, PhD

Metrics, Morbidity, and Mortality Among the Elderly

Principal Investigator: Ari Friedman, MD, PhD | Perelman School of Medicine

In this study, we performed a survey-weighted analysis of the National Hospital Ambulatory Medical Care Survey (NHAMCS), comparing older adults with a chief complaint of abdominal pain to those without. Visits from 2013 to 2017 to nationally representative emergency departments (ED) were included. We analyzed 81,509 visits to 1211 U.S. EDs, which projects to 531,780,629 ED visits after survey weighting. We report the diagnostic testing, evaluation, management, additional reasons for visit, and disposition of ED visits. We found that among older adults (≥65 years), seven% of ED visits were for abdominal pain. Older patients with abdominal pain had a lower probability of being triaged to the “Emergent” (ESI2) acuity on arrival (7.1% vs. 14.8%), yet were more likely to be admitted directly to the operating room than older adults without abdominal pain (3.6% vs. 0.8%), with no statistically significant differences in discharge home, death, or admission to critical care. Ultrasound or CT imaging was performed in 60% of older adults with abdominal pain. A minority (39%) of older patients with abdominal pain received an electrocardiogram (EKG).

Resulting Product:


James Guevara
James Guevara, MD, MPH

Literacy Promotion Using Automated Hovering to Improve Development in Young Low-Income Children

Principal Investigator: James Guevara, MD, MPH | Perelman School of Medicine

We conducted a mixed-methods study consisting of rapid cycle interviews and a pilot RCT to determine if automated hovering approaches can increase parent-child shared reading. Rapid cycle interviews suggested the importance of daily reading routines, reciprocal engagement in reading, reliance on past experiences, contribution of reading to language development, challenges in habitual reading, and reminder notifications. RCT results suggest that the regret arm had the greatest improvement in the home reading environment scores and the least worsening in parenting stress scores. Responses to text queries on daily reading were inconsistent.


Atul Gupta
Atul Gupta, PhD

How Do Health Care Providers Deploy Public Funds?

Principal Investigator: Atul Gupta, PhD | The Wharton School

The potential detrimental impact of a “Medicare for All” system on health care provider finances is an important policy debate concern. This study focuses on a related setting — the recent public insurance expansions authorized by the Affordable Care Act (ACA) that led to a decline in uncompensated care burden for providers. The goal of this study is to quantify the deployment of the incremental funds towards various aspects of care production and other relevant activities.


Eileen Lake
Eileen Lake, PhD, RN

Disparities in Neonatal Abstinence Syndrome Infant Outcomes and Related Nursing System Factors

Principal Investigator: Eileen Lake, PhD, RN | Penn Nursing and School of Arts & Sciences

This study used secondary data analysis of hospital discharge abstracts and RN4CAST-US nurse survey data from three states (California, New Jersey, and Pennsylvania) in 2016 to explore the extent of segregation and care inequality among infants with neonatal abstinence syndrome (NAS). We found that NAS infants (356 Black [9.05%] and 3,578 white [90.95%]) were segregated among the 202 hospitals by race. The hospital segregation index for black NAS infants was 0.106. Black and white NAS infants received care in 77 hospitals with different quality/safety ratings. Among the 77 hospitals, fewer Black NAS infants received care in hospitals with higher quality/safety ratings compared white NAS infants. In lower quality hospitals, there were more Black infants than white infants. Infants with NAS are racially segregated across newborn care settings to a modest extent. Inequality of care is evident in this patient population in a manner similar to that seen in neonatal intensive care units. We need policy that allows for the measurement of quality in this population and we need policy to allocate resources to shore up the quality of care in poorer performing hospitals.

Resulting Products:


Terri Lipman
Terri Lipman, PhD, CRNP

Impact of Community Health Workers in High Risk Pediatric Patients With Newly Diagnosed Type 1 Diabetes

Principal Investigator: Terri Lipman, PhD, CRNP | Penn Nursing

Adverse social determinants of health (SDOH) were identified by 80% of the intervention group at baseline and 50% in the control group. Our preliminary data demonstrated that adding a community health worker to the care of these children reduced the prevalence of adverse SDOH, when compared with controls. There was a marked decrease of SDOH from 80% to 25% at 12 months in the intervention group. Caregivers reported having difficulty adjusting to new routines and care practices after discharge. Support systems made this transition somewhat easier, particularly having access to community health workers.

Resulting Products:


David Mandell
David Mandell, ScD

Advancing Behavior Change Among Older Adults: When and Why Financial Incentives Succeed or Fail

Principal Investigator: David Mandell, ScD | Perelman School of Medicine and Children’s Hospital of Philadelphia

This pilot study was designed to test the hypothesis that the strength of motivation to change a behavior predicts the likelihood of accepting the invitation to enroll in a financially incentivized behavior change study. To measure motivation, we adapted a validated measure of intention, which captures the perceived likelihood that one will exercise more or sleep more in the future. The participants were then randomized to an invitation to join a study designed to increase either sleep or physical activity. This study was described as providing financial incentives to those who succeed in changing their behavior. In our study sample, the strength of motivation or intention to increase sleep and exercise varied. We found a statistically significant association between strength of motivation to increase sleep and exercise, and whether the participant enrolled in a randomized trial of incentives to change that behavior. Age did not moderate the association.


Stephanie Mayne
Stephanie Mayne, PhD, MHS

The Get to Sleep Study: Assessing the Feasibility of Using Mobile Methods to Assess Neighborhood Context and Sleep Among Urban Adolescents

Principal Investigator: Stephanie Mayne, PhD, MHS | Children’s Hospital of Philadelphia

We enrolled a convenience sample of 25 adolescents aged 15-17 (mean age: 16, 44% female) through CHOP’s Recruitment Enhancement Core (REC). We found that our seven-day protocol of GPS tracking, ecological momentary assessment, and sleep tracking was feasible and acceptable to participants. Most teens had GPS data available on all seven days. Participants’ mean sleep duration was 7.2 hours. Seventy-eight percent of EMA surveys were completed overall (median across teens: 86%, range 29-100%). Nearly all participants (96%) reported no privacy concerns about GPS tracking and that EMA survey burden was minimal.


Ashley Ritter
Ashley Ritter, PhD, CRNP

Prevalence and Patterns of Opioid Use Disorder among Veterans Discharged to Post-Acute Care

Principal Investigator: Ashley Ritter, PhD, CRNP | NewCourtland

Using Medicare and VA data, we identified 25,871 veterans over age 65 with a diagnosis of opioid use disorder (OUD) who were discharged to a post-acute care (PAC) setting between 2012 and 2016. PAC destinations were relatively stable over time, though patterns differed between veterans with and without a diagnosis of OUD. Veterans over 65 with a diagnosis of OUD are less often referred to home health services when compared to those who do not have a diagnosis of OUD. Use of nursing home care was similar between the two groups. The remaining discharges among veterans with OUD went to VA PAC settings (Community Living Centers and State Veteran Homes). More veterans with a diagnosis of OUD experienced more than one PAC stay during the study period (56.5%) when compared to veterans without a diagnosis of OUD (43.6%). Veterans experiencing OUD discharged to PAC were younger and sicker, as evidenced by higher Charlson scores, longer hospital length of stays, and more frequent ICU stays on average. Veterans over 65 with a diagnosis of OUD experienced a much higher rate of rehospitalization (24.59% vs. 10.87%) with lower rate of death by 100 days (10.65% vs. 14.52%) when compared to older veterans without a diagnosis of OUD. Our findings demonstrate older veterans with a diagnosis of OUD utilize nursing home care following hospitalization, with a growing proportion discharged to VA nursing homes. In addition, older veterans discharged to PAC settings experience health care readmissions. Interventions should target the specific needs of veterans in PAC settings to reduce costly and potentially burdensome readmissions.


Brittany Rudd
Brittany Rudd, PhD

Pilot Mixed Methods Evaluation of the Barriers and Facilitators to Suicide Prevention in Juvenile Detention

Principal Investigator: Brittany Rudd, PhD | The University of Illinois at Chicago

We sought to identify the barriers to and facilitators of implementing the Zero Suicide model in juvenile detention. Survey results revealed that leadership support for suicide prevention (SP) and volume and quality of SP training positively related to engagement in SP. Qualitative results highlighted barriers to effective suicide prevention, including systemic obstacles in youth environment, lack of behavioral health care, and organizational staffing processes. Facilitators included existence of codified SP practices, avenues of communication, and a trauma-informed culture. Results will inform the development of an implementation strategy.

Resulting Products:


Aditi Vasan
Aditi Vasan, MD, MSHP

Pediatric Caregiver Perspectives Regarding Use of an Electronic Resource Map to Address Unmet Social Needs

Principal Investigator: Aditi Vasan, MD, MSHP | Perelman School of Medicine

In this pilot study, we aimed to examine pediatric caregiver perspectives regarding the feasibility and acceptability of using an electronic resource map to identify available community resources and to identify barriers to and facilitators of connection to these resources among pediatric caregivers with unmet social needs. From July 2020 onwards, we screened 193 caregivers for unmet social needs, and 72 (37%) screened positive for one or more social needs. The most prevalent needs in our population were depressed mood (23% of caregivers), food insecurity (19%), and need for assistance with utilities (10%). We conducted follow-up surveys with 125 caregivers and follow-up interviews with 28 caregivers, exploring barriers to and facilitators of resource connection. Caregivers who completed our follow-up surveys and interviews were predominantly Medicaid-insured (82%) and had children with one or more complex chronic conditions (68%). We identified four primary themes from our qualitative analysis. First, caregivers described competing priorities related to caring for a medically complex child as a barrier to resource connection. Second, caregivers described administrative burdens, like complex application and enrollment processes, as a barrier to applying for and utilizing government benefit programs like WIC, SNAP, and LIHEAP, particularly during the COVID-19 pandemic. Third, caregivers expressed appreciation for electronic resources and information about locally available resources as a facilitator of resource connection. Lastly, many caregivers expressed a desire for longitudinal support in establishing and maintaining connection with resources after hospital discharge.

Resulting Products: