Beth W Karlson

Dr. Beth W Karlson


Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
NBNatalie Boutin
Natalie Boutin
SSSamantha Schecter
Samantha Schecter
EPEmma F. Perez
Emma F. Perez
NTNatasha Tchamitchian
Natasha Tchamitchian
XCXander Cerretani
Xander Cerretani
JSJordan Smoller
Jordan Smoller
The Mass General Brigham Biobank (formerly Partners HealthCare Biobank) is a large repository of biospecimens and data linked to extensive electronic health record data and survey data. Its objective is to support and enable translational research focused on genomic, environmental, biomarker and family history associations with disease phenotypes. The Biobank has enrolled more than 135,000 participants, generated genomic data on more than 65,000 of its participants, distributed approximately 153,000 biospecimens, and served close to 450 institutional studies with biospecimens or data. Although the Biobank has been successful, based on some measures of output, this has required substantial institutional investment. In addition, several challenges are ongoing, including: (1) developing a sustainable cost model that doesn’t rely as heavily on institutional funding; (2) integrating Biobank operations into clinical workflows; and (3) building a research resource that is diverse and promotes equity in research. Here, we describe the evolution of the Biobank and highlight key lessons learned that may inform other efforts to build biobanking efforts in health system contexts.

29 May 2018 in BMC Medical Informatics and Decision Making
Verified
Journal Article
Open access
PDF file
QZQiu-Yue Zhong
Qiu-Yue Zhong
BKBeth W Karlson
Beth W Karlson
BGBizu Gelaye
Bizu Gelaye
SFSean Finan
Sean Finan
PAPaul Avillach
Paul Avillach
MWMichelle A. Williams
Michelle A. Williams
We examined the comparative performance of structured, diagnostic codes vs. natural language processing (NLP) of unstructured text for screening suicidal behavior among pregnant women in electronic medical records (EMRs). Women aged 10–64 years with at least one diagnostic code related to pregnancy or delivery (N = 275,843) from Partners HealthCare were included as our “datamart.” Diagnostic codes related to suicidal behavior were applied to the datamart to screen women for suicidal behavior. Among women without any diagnostic codes related to suicidal behavior (n = 273,410), 5880 women were randomly sampled, of whom 1120 had at least one mention of terms related to suicidal behavior in clinical notes. NLP was then used to process clinical notes for the 1120 women. Chart reviews were performed for subsamples of women. Using diagnostic codes, 196 pregnant women were screened positive for suicidal behavior, among whom 149 (76%) had confirmed suicidal behavior by chart review. Using NLP among those without diagnostic codes, 486 pregnant women were screened positive for suicidal behavior, among whom 146 (30%) had confirmed suicidal behavior by chart review. The use of NLP substantially improves the sensitivity of screening suicidal behavior in EMRs. However, the prevalence of confirmed suicidal behavior was lower among women who did not have diagnostic codes for suicidal behavior but screened positive by NLP. NLP should be used together with diagnostic codes for future EMR-based phenotyping studies for suicidal behavior.

5 January 2019 in BMC Musculoskeletal Disorders
Verified
Journal Article
Open access
PDF file
JSJeffrey A. Sparks
Jeffrey A. Sparks
ÉOÉilis J. O’rEilly
Éilis J. O’rEilly
MBMedha Barbhaiya
Medha Barbhaiya
Sara K. TedeschiSara K. Tedeschi
Sara K. Tedeschi
SMSusan Malspeis
Susan Malspeis
BKBeth W Karlson
Beth W Karlson
Prior studies suggest that fish may be protective for rheumatoid arthritis (RA) risk perhaps through the anti-inflammatory effect of omega-3 fatty acid, but this relationship has not been clearly established. Therefore, we investigated fish intake and RA risk by serologic status, age of onset, and smoking using a prospective cohort study with large sample size, repeated measures of dietary intake, and lengthy follow-up. We studied fish intake and RA risk among 166,013 women in two prospective cohorts, the Nurses’ Health Study (NHS, 1984–2014) and NHSII (1991–2015). Fish intake was assessed using food frequency questionnaires at baseline and every 4 years. Incident RA during follow-up and serologic status were determined by medical record review. Pooled Cox regression models estimated hazard ratios (HR) and 95% confidence intervals (CI) for RA (overall and by serologic status and age at diagnosis) for fish intake frequency. We tested for a smoking-fish interaction for RA risk. During 3,863,909 person-years of follow-up, we identified 1080 incident RA cases. Increasing fish intake was not associated with all RA (≥4 servings/week: multivariable HR 0.93 [95%CI 0.67–1.28] vs. < 1 serving/month; p for trend = 0.42), seropositive RA (p for trend = 0.66), or seronegative RA (p for trend = 0.45), but had increased risk for RA diagnosed > 55 years old (p for trend = 0.037). Among women ≤55 years old, frequent fish intake (vs. infrequent) had HRs (95%CIs) of: 0.73 (0.52–1.02) for all RA, 0.85 (0.55–1.32) for seropositive RA, and 0.55 (0.32–0.94) for seronegative RA. Ever smokers with infrequent fish intake had highly elevated risk for RA onset ≤55 years (HR 2.59, 95%CI 1.65–4.06), while ever smokers with frequent fish intake had modestly elevated RA risk (HR 1.29, 95%CI 1.07–1.57; vs. never smokers/frequent fish intake; p for smoking-fish interaction = 0.039). In this large prospective cohort study, we found no clear protective effect of fish or marine omega-3 fatty acid intake on RA risk, overall or by serologic status. We found that fish intake attenuated the strong association of smoking for RA diagnosed ≤55 years of age, but this requires further study.

1 May 2019 in Patient Education and Counseling
Verified
Journal Article
Open access
AMAllison A. Marshall
Allison A. Marshall
AZAlessandra Zaccardelli
Alessandra Zaccardelli
ZYZhi Yu
Zhi Yu
MPMaria G. Prado
Maria G. Prado
XLXinyi Liu
Xinyi Liu
JSJeffrey A. Sparks
Jeffrey A. Sparks
To investigate the effect of providing comprehensive personalized risk information on concern for chronic disease development. Unaffected first-degree relatives (FDRs) of rheumatoid arthritis (RA) patients (n = 238) were randomly allocated to: 1) disclosure of RA risk personalized to demographics, genetics, biomarkers, and behaviors using a web-based tool (PRE-RA arm, n = 78); 2) PRE-RA with interpretation by a health educator (PRE-RA Plus arm, n = 80); and 3) standard RA education (Comparison arm, n = 80). Concern for developing RA was assessed at baseline and immediately, 6 weeks, 6 months, and 12 months post-intervention. FDRs randomized to PRE-RA arms were less concerned about developing RA than the Comparison arm at all post-intervention assessments (p < 0.05). Among those concerned about RA risk at baseline, the PRE-RA (OR = 4.7, 95%CI 1.5-14.4) and PRE-RA Plus (OR = 5.2, 95%CI 1.6-17.3) arms were more likely to have reassurance 6 months post-intervention than the Comparison arm. A comprehensive tool provided reassurance to those at risk for developing a chronic disease, with or without interpretation from a health educator, compared to standard education. Individuals may be more likely to be reassured using a personalized chronic disease risk disclosure tool than a standard non-personalized approach.

1 July 2021 in The Journal of Applied Laboratory Medicine
Verified
Journal Article
Open access
PDF file
ENEric J Nilles
Eric J Nilles
BKBeth W Karlson
Beth W Karlson
MNMaia Norman
Maia Norman
TGTal Gilboa
Tal Gilboa
SFStephanie Fischinger
Stephanie Fischinger
LBLindsey R Baden
Lindsey R Baden
Background: Serological testing provides a record of prior infection with SARS-CoV-2, but assay performance requires independent assessment. Methods: We evaluated 3 commercial (Roche Diagnostics pan-IG, and Epitope Diagnostics IgM and IgG) and 2 non-commercial (Simoa and Ragon/MGH IgG) immunoassays against 1083 unique samples that included 251 PCR-positive and 832 prepandemic samples. Results: The Roche assay registered the highest specificity 99.6% (3/832 false positives), the Ragon/MGH assay 99.5% (4/832), the primary Simoa assay model 99.0% (8/832), and the Epitope IgG and IgM 99.0% (8/830) and 99.5% (4/830), respectively. Overall sensitivities for the Simoa, Roche pan-IG, Epitope IgG, Ragon/MGH IgG, and Epitope IgM were 92.0%, 82.9%, 82.5%, 64.5% and 47.0%, respectively. The Simoa immunoassay demonstrated the highest sensitivity among samples stratified by days postsymptom onset (PSO), 21 days PSO (95.18%). Conclusions: All assays demonstrated high to very high specificities while sensitivities were variable across assays.

7 October 2021 in Scientific Reports
Verified
Journal Article
Open access
PDF file
SCSu Chu
Su Chu
EWEmily S. Wan
Emily S. Wan
MCMichael H. Cho
Michael H. Cho
SGSergey Goryachev
Sergey Goryachev
VGVivian Gainer
Vivian Gainer
BKBeth W Karlson
Beth W Karlson
Electronic health records (EHR) provide an unprecedented opportunity to conduct large, cost-efficient, population-based studies. However, the studies of heterogeneous diseases, such as chronic obstructive pulmonary disease (COPD), often require labor-intensive clinical review and testing, limiting widespread use of these important resources. To develop a generalizable and efficient method for accurate identification of large COPD cohorts in EHRs, a COPD datamart was developed from 3420 participants meeting inclusion criteria in the Mass General Brigham Biobank. Training and test sets were selected and labeled with gold-standard COPD classifications obtained from chart review by pulmonologists. Multiple classes of algorithms were built utilizing both structured (e.g. ICD codes) and unstructured (e.g. medical notes) data via elastic net regression. Models explicitly including and excluding spirometry features were compared. External validation of the final algorithm was conducted in an independent biobank with a different EHR system. The final COPD classification model demonstrated excellent positive predictive value (PPV; 91.7%), sensitivity (71.7%), and specificity (94.4%). This algorithm performed well not only within the MGBB, but also demonstrated similar or improved classification performance in an independent biobank (PPV 93.5%, sensitivity 61.4%, specificity 90%). Ancillary comparisons showed that the classification model built including a binary feature for FEV1/FVC produced substantially higher sensitivity than those excluding. This study fills a gap in COPD research involving population-based EHRs, providing an important resource for the rapid, automated classification of COPD cases that is both cost-efficient and requires minimal information from unstructured medical records.