Privacy-preserving record linkage across disparate institutions and datasets to enable a learning health system: The national COVID cohort collaborative (N3C) experience

Introduction

Research driven by real-world clinical data is increasingly vital to enabling learning health systems, but integrating such data from across disparate health systems is challenging. As part of the NCATS National COVID Cohort Collaborative (N3C), the N3C Data Enclave was established as a centralized repository of deidentified and harmonized COVID-19 patient data from institutions across the US. However, making this data most useful for research requires linking it with information such as mortality data, images, and viral variants. The objective of this project was to establish privacy-preserving record linkage (PPRL) methods to ensure that patient-level EHR data remains secure and private when governance-approved linkages with other datasets occur.

Methods

Separate agreements and approval processes govern N3C data contribution and data access. The Linkage Honest Broker (LHB), an independent neutral party (the Regenstrief Institute), ensures data linkages are robust and secure by adding an extra layer of separation between protected health information and clinical data. The LHB's PPRL methods (including algorithms, processes, and governance) match patient records using "deidentified tokens," which are hashed combinations of identifier fields that define a match across data repositories without using patients' clear-text identifiers.

Results

These methods enable three linkage functions: Deduplication, Linking Multiple Datasets, and Cohort Discovery. To date, two external repositories have been cross-linked. As of March 1, 2023, 43 sites have signed the LHB Agreement; 35 sites have sent tokens generated for 9 528 998 patients. In this initial cohort, the LHB identified 135 037 matches and 68 596 duplicates.

Conclusion

This large-scale linkage study using deidentified datasets of varying characteristics established secure methods for protecting the privacy of N3C patient data when linked for research purposes. This technology has potential for use with registries for other diseases and conditions.

Sex, obesity, diabetes, and exposure to particulate matter among patients with severe asthma: Scientific insights from a comparative analysis of open clinical data sources during a five-day hackathon

This special communication describes activities, products, and lessons learned from a recent hackathon that was funded by the National Center for Advancing Translational Sciences via the Biomedical Data Translator program ('Translator'). Specifically, Translator team members self-organized and worked together to conceptualize and execute, over a five-day period, a multi-institutional clinical research study that aimed to examine, using open clinical data sources, relationships between sex, obesity, diabetes, and exposure to airborne fine particulate matter among patients with severe asthma. The goal was to develop a proof of concept that this new model of collaboration and data sharing could effectively produce meaningful scientific results and generate new scientific hypotheses. Three Translator Clinical Knowledge Sources, each of which provides open access (via Application Programming Interfaces) to data derived from the electronic health record systems of major academic institutions, served as the source of study data. Jupyter Python notebooks, shared in GitHub repositories, were used to call the knowledge sources and analyze and integrate the results. The results replicated established or suspected relationships between sex, obesity, diabetes, exposure to airborne fine particulate matter, and severe asthma. In addition, the results demonstrated specific differences across the three Translator Clinical Knowledge Sources, suggesting cohort- and/or environment-specific factors related to the services themselves or the catchment area from which each service derives patient data. Collectively, this special communication demonstrates the power and utility of intense, team-oriented hackathons and offers general technical, organizational, and scientific lessons learned.

Meaningful use of electronic behavioral health data in primary health care

In August 2011, scientists and policy-makers held a conference entitled "Using IT to Improve Community Health: How Health Care Reform Supports Innovation." One of the conference sessions was entitled "Electronic health records: Meaningful use implementation challenges, innovation, and regulations." This Meeting Report discusses the meaningful use of behavioral health data for the treatment of mental health and substance abuse conditions and optimization of behavioral wellness by primary care physicians.

Support for the vascular depression hypothesis in late-life depression: results of a 2-site, prospective, antidepressant treatment trial

CONTEXT

Research on vascular depression has used 2 approaches to subtype late-life depression, based on executive dysfunction or white matter hyperintensity severity.

OBJECTIVE

To evaluate the relationship of neuropsychological performance and white matter hyperintensity with clinical response in late-life depression.

DESIGN

Two-site, prospective, nonrandomized controlled trial.

SETTING

Outpatient clinics at Washington University and Duke University.

PARTICIPANTS

A total of 217 subjects aged 60 years or older met DSM-IV criteria for major depression, scored 20 or more on the Montgomery-Asberg Depression Rating Scale (MADRS), and received vascular risk factor scores, neuropsychological testing, and magnetic resonance imaging; they were excluded for cognitive impairment or severe medical disorders. Fazekas rating was conducted to grade white matter hyperintensity lesions. Intervention Twelve weeks of sertraline treatment, titrated by clinical response. Main Outcome Measure Participants' MADRS scores over time.

RESULTS

Baseline neuropsychological factor scores correlated negatively with baseline Fazekas scores. A mixed model examined effects of predictor variables on MADRS scores over time. Baseline episodic memory (P = .002), language (P = .007), working memory (P = .01), processing speed (P < .001), executive function factor scores (P = .002), and categorical Fazekas ratings (P = .05) predicted MADRS scores, controlling for age, education, age of onset, and race. Controlling for baseline MADRS scores, these factors remained significant predictors of decrease in MADRS scores, except for working memory and Fazekas ratings. Thirty-three percent of subjects achieved remission (MADRS < or =7). Remitters differed from nonremitters in baseline cognitive processing speed, executive function, language, episodic memory, and vascular risk factor scores.

CONCLUSIONS

Comprehensive neuropsychological function and white matter hyperintensity severity predicted MADRS scores prospectively over a 12-week treatment course with selective serotonin reuptake inhibitors in late-life depression. Baseline neuropsychological function differentiated remitters from nonremitters and predicted time to remission in a proportional hazards model. Predictor variables correlated highly with vascular risk factor severity. These data support the vascular depression hypothesis and highlight the importance of linking subtypes based on neuropsychological function and white matter integrity.

TRIAL REGISTRATION

clinicaltrials.gov Identifier: NCT00045773.

Clinical characteristics and resource utilization of patients with bipolar disorder who have frequent psychiatric interventions

OBJECTIVE

To compare the demographics, clinical characteristics and resource utilization of patients with bipolar disorder who required frequent psychiatric interventions (FPIs) with those needing fewer interventions in the Duke Healthcare System database between 1999 and 2005.

METHODS

This retrospective analysis was conducted using electronic medical records of bipolar patients with FPIs, defined as having ≥4 clinically significant events (CSEs) in any 12-month period while in the Duke University Healthcare System. CSEs were composed of emergency room visits, inpatient hospitalizations, or a change in psychotropic medication due to psychiatric symptoms (score≥4 on the Clinical Global Impressions-Severity scale). Data were compared between patients with and without FPIs.

RESULTS

Of 632 patients with bipolar disorder 52.5% were identified as having FPIs. These patients were younger and more often female and African American than those with fewer interventions (p<0.01 for all). Patients with FPIs were generally prescribed more psychotropic and non-psychotropic medications, utilized more healthcare resources and experienced more psychiatric co-morbidities than those who did not require FPIs (p<0.01 for all).

LIMITATIONS

These results are from a single healthcare system and may not be generalizable to all patients with bipolar disorder. This analysis was retrospective and relied on availability of adequate information recording and coding of diagnoses by physicians.

CONCLUSIONS

Patients with bipolar disorder who required FPIs were significantly different from those with fewer clinically defined interventions with respect to their demographic and clinical characteristics and prescribed medications.

Patterns of pharmacotherapy and treatment response in elderly adults with bipolar disorder

BACKGROUND

Bipolar disorder is a difficult disease to treat because of its cycling nature, frequent residual symptoms, and poor compliance to treatment. Several guidelines have been proposed for treatment, but there is limited data on best treatment practices in elderly, bipolar patients. This study assessed patterns of psychopharmacological treatment and treatment response in acutely ill, bipolar patients over the age of 60.

METHODS

Naturalistic pharmacologic data was obtained on 138 acutely ill, elderly bipolar patients from the Duke University Medical Center electronic psychiatric record. Clinical Global Impression (CGI) severity and improvement scale ratings were performed at each visit, and time to response was measured. Pharmacological treatment selected was reviewed.

RESULTS

Standard mood stabilizers (lithium, valproate, carbamazepine, and lamotrigine) were the most prescribed medications (68%), followed by antipsychotics (54%) and antidepressants (34%). Combination therapy was more common than monotherapy (57% vs. 38%). Remission was achieved in 35% of subjects, while 32% showed no significant improvement. There was no difference in antipsychotic prescription between old old and young old patients.

CONCLUSIONS

In this naturalistic, real-setting study of pharmacologic treatment, acutely ill, elderly bipolar patients were treated primarily with mood-stabilizing agents, followed by antipsychotics and antidepressants. Combination therapy is much more common than monotherapy. Results can be useful in understanding the current clinical standard of care in elderly bipolar patients, and are consistent with current clinical guidelines for mixed-age bipolar patients.

Effect of bupropion extended release on negative emotion processing in major depressive disorder: a pilot functional magnetic resonance imaging study

BACKGROUND

Prior imaging studies suggest that patients with major depressive disorder have abnormalities in frontal and limbic neural circuitry including the amygdala, which is relatively more activated at rest and in response to negative emotional stimuli (sadness, fear, etc.) in depressed patients than in controls. Concurrently, patients with depression may have decreased activation of attentional executive regions in response to attentional stimuli. This study examined the effect of bupropion XL, an extended release formulation of the nonserotonergic antidepressant agent bupropion, using a paradigm that investigated both negative emotional response and attentional processing.

METHOD

Functional magnetic resonance imaging (fMRI) scans and clinical ratings were obtained for 10 patients with DSM-IV-TR-defined major depressive disorder (mean [SD] age = 41 [+/- 7] years, mean [SD] Hamilton Rating Scale for Depression [HAM-D] score = 21 [+/- 4]) before and after 8 weeks of treatment with bupropion XL. The fMRI sessions were conducted during administration of the Emotional Oddball Task; scans were obtained while subjects viewed emotional distracters and performed an attentional executive function task. The primary outcome was fMRI activations evoked by the emotional distracters. The first baseline fMRI scan was performed in December 2004, and the last posttreatment scan was in March 2005.

RESULTS

Treatment with bupropion XL was associated with improvements in HAM-D and Clinical Global Impressions scale ratings (p < .05). Treatment reduced fMRI activation during emotional distracters in several regions including right orbital frontal cortex, left dorsomedial prefrontal cortex, right ventromedial prefrontal cortex, right anterior cingulate cortex, right inferior frontal cortex, right amygdala/parahippocampal area, right caudate, right fusiform gyrus, and left posterior cingulate. In addition, changes in fMRI activation in the amygdala correlated with improvements on the HAM-D (p < .05). Treatment increased activation to attentional targets in the following regions: right middle and inferior frontal gyri, right caudate, and bilateral precuneus.

CONCLUSION

Despite the limitations of a small sample size and the lack of a placebo control group, this study demonstrated that bupropion XL therapy for 8 weeks may attenuate emotion-induced, blood-oxygen-level-dependent (BOLD) activation responses in the amygdala and related brain regions. Such attenuation may be associated with a positive clinical response in depression. Bupropion XL also improved activation in the executive-function neural network. These fMRI surrogate markers offer promise for studying antidepressant and neurocognitive effects of existing and novel therapies.

Effect of mirtazapine orally disintegrating tablets on health-related quality of life in elderly depressed patients with comorbid medical disorders: a pilot study

BACKGROUND

There is a need for additional studies on the quality of life (QOL) of elderly depressed subjects with medical comorbidity.

METHOD

We conducted a 10-week, open trial of mirtazapine orally disintegrating tablets in 16 elderly subjects with major depressive disorder and one or more serious medical illnesses. Quality of life was measured by the Medical Outcomes Study Short Form-36 Health Status Survey (SF- 36).

RESULTS

Treatment with mirtazapine was associated with significant reductions in clinical global impressions-severity of illness scale (CGI-S) score, the Hamilton rating scale for anxiety (HAM-A) total score, the 17-item Hamilton rating scale for depression (HAM-D) total score and the Beck depression inventory (BDI) total scores. The SF-36 "physical functioning", "role limitation physical", "vitality", "social functioning", "role limitation emotional", and "mental health" domains improved significantly. The mean mirtazapine dose at endpoint was 35 mg per day. The drug was relatively well tolerated except for three subjects who dropped out because of side effects. No drug-drug interactions or significant changes in blood pressure or heart rate occurred.

CONCLUSION

Mirtazapine orally disintegrating tablets may improve depression, insomnia, anxiety, somatic symptoms, and certain quality-of-life measures in elderly depressed subjects with medical disorders. A randomized, placebo-controlled study is warranted to confirm these promising findings.

Cognitive function in late life depression: relationships to depression severity, cerebrovascular risk factors and processing speed

BACKGROUND

A number of studies have examined clinical factors linked to worse neuropsychological performance in late life depression (LLD). To understand the influence of LLD on cognition, it is important to determine if deficits in a number of cognitive domains are relatively independent, or mediated by depression- related deficits in a basic domain such as processing speed.

METHODS

Patients who met DSM-IV criteria for major depression (n = 155) were administered a comprehensive neuropsychological battery of tasks grouped into episodic memory, language, working memory, executive function, and processing speed domains. Multiple regression analyses were conducted to determine contributions of predictor variables to cognitive domains.

RESULTS

Age, depression severity, education, race and vascular risk factors all made significant and independent contributions to one or more domains of cognitive function, with all five making independent contributions to processing speed. Age of onset made no independent contribution, after accounting for age and vascular risk factors. Of the five cognitive domains investigated, changes in processing speed were found to most fully mediate the influence of predictor variables on all other cognitive domains.

CONCLUSIONS

While slowed processing speed appears to be the most core cognitive deficit in LLD, it was closely followed by executive function as a core cognitive deficit. Future research is needed to help clarify mechanisms leading to LLD- related changes in processing speed, including the potential role of white matter abnormalities.

Medical comorbidity in a bipolar outpatient clinical population

The presence of medical illnesses among inpatients with bipolar disorder is known to complicate treatment and lengthen hospital stay. However, except for a few specific diseases, little is known about prevalence of medical illnesses in bipolar outpatients and the effect it may have on treatment. The authors sought to assess the presence of medical illnesses in a large outpatient clinical sample of bipolar patients, and the effect that medical illnesses may have on the clinical assessment and treatment of the underlying bipolar disorder. Using the Duke University Medical Center clinical database, the authors categorized the medical diagnoses of 1379 patients who were treated with bipolar disorder from 2001 to 2002 through outpatient psychiatric clinics. The prevalence of medical comorbidities was examined, as well as the effect their presence had on the clinician's assessment of disease severity and time to improvement. As expected, medical comorbidities increased with age. The most common systemic illnesses in bipolar outpatients were Endocrine and Metabolic Diseases (13.6% of the sample), Diseases of the Circulatory System (13.0%), and Diseases of the Nervous System and Sense Organs (10.7%). Significant specific diseases included cardiovascular diseases/hypertension (10.7%), COPD/asthma (6.1%), diabetes (4.3%), HIV infection (2.8%), and hepatitis C infection (1.9%). Clinicians assessed greater severity of illness in patients with increasing numbers of comorbid conditions; however, the time to recovery was not significantly effected by the presence of medical comorbidity. In conclusion, comorbid medical illnesses are common in bipolar outpatients, increasing with age. HIV rates may be increased relative to population norms. Their presence compounds the severity of the illness at time of presentation.

Informant-rated cognitive symptoms in normal aging, mild cognitive impairment, and dementia. Initial development of an informant-rated screen (Brief Cognitive Scale) for mild cognitive impairment and dementia

An informant-rated cognitive screen may have the potential to reliably help detect early dementia. A valuable scale should have good interitem associations and strong reliability when tested in groups with and without cognitive impairment. Our scale, the Brief Cognitive Scale (BCS), consists of 18 questions designed to assess cognitive function that affects everyday activities. Each question is coded with one of four levels, ranging from no impairment to severe impairment. We administered this screen to 120 subjects: 26 controls, 28 with a diagnosis of mild cognitive impairment (MCI), and 66 with a diagnosis of dementia. In addition, we administered a Folstein Mini-Mental Status Examination (MMSE) to each subject. Our results showed that the BCS scores were lowest in the control group and highest in the dementia group. In our sample, this scale was effective at discriminating between subjects with no cognitive impairment, MCI, and dementia. However, the scale needs further refinement before it can be employed in a clinical setting.