Understanding challenges of using routinely collected health data to address clinical care gaps: a case study in Alberta, Canada

The effect of warm ischemic intervals on primary graft dysfunction in normothermic regional perfusion for donation after circulatory death heart transplant

OBJECTIVES

To clarify the association between warm ischemic time during donation after circulatory death (DCD) and severe primary graft dysfunction (PGD) after heart transplant.

METHODS

DCD heart transplants using normothermic regional perfusion, excluding congenital etiology or multiorgan transplant, at a single institution from January 2020 to December 2024 were reviewed. Donation withdrawal ischemic time (DWIT), functional warm ischemic time, defined by oxygen saturation <80% (FWIT O2), systolic blood pressure <80 mm Hg or <50 mm Hg, and asystolic warm ischemic time were examined. Propensity matching created balanced cohorts to associate warm ischemia and outcomes. Outcomes included incidence of severe PGD, lengths of stay, and mortality.

RESULTS

The final study cohort had 135 patients, of whom 10 of 135 (7.4%) had severe PGD. When stratified by severe PGD, donor and recipient demographics were similar. DWIT (median 25.0 minutes vs 35.5 minutes, P = .031) and FWIT O2 (median 22.0 vs 33.0 minutes, P = .025) were lower in those without severe PGD. Logistic regression identified FWIT O2 as a better predictor compared with DWIT. Receiver operating characteristic curve analysis identified a FWIT threshold of 23 minutes (area under the curve, 0.714). After matching, rates of severe PGD were significantly greater in the FWIT O2 >23 minutes group (8/59 [13.6%] vs 1/59 [1.7%], P = .032). However, the FWIT O2 >23 minutes group had similar lengths of stay and mortality.

CONCLUSIONS

In DCD normothermic regional perfusion heart transplant, >23 minutes of FWIT O2 is associated with increased rates of severe PGD. However, increased FWIT O2 was not associated with other outcomes, including mortality. Rejection of allografts on the basis of prolonged warm ischemia may lead to unnecessary discard of viable hearts.

Individuals with Cognitive Impairment Entering Long-Term Care: Characteristics and Cumulative Incidence of Dementia after Care Entry

OBJECTIVES

To characterize individuals entering long-term care facilities (LTCFs) with evidence of cognitive impairment and without a diagnosis of dementia, and to ascertain the cumulative incidence of dementia after care entry.

DESIGN

Retrospective cohort study using the Registry of Senior Australians (ROSA) National Historical Cohort.

SETTING AND PARTICIPANTS

Individuals aged 65 to 105 years who entered LTCFs between 2009 and 2018, received a cognitive evaluation, and had no recorded dementia diagnosis at the time of care entry.

METHODS

Cognitive function was determined via the Psychogeriatric Assessment Scales-Cognitive Impairment Scales (PAS-CIS) and defined as none or minimal (PAS-CIS score 0 to <4), mild (4 to <10), or moderate to severe (10 to 21). The cumulative incidence of dementia, determined by aged care assessments, hospitalization, medication, or cause of death, was ascertained for the total cohort and by cognitive impairment status at care entry.

RESULTS

In total, 90,122 individuals [median age 85 years; interquartile range (IQR) 81-89; 64.6% female] were studied, of whom 76.6% (n = 69,075) had cognitive impairment, including 51.4% (n = 46,350) with mild and 25.2% (n = 22,725) with moderate to severe impairment. Over a median follow-up of 1.5 years (IQR 0.6-2.9), the cumulative incidence of dementia was 26.8% [95% confidence interval (CI), 26.5-27.1]. Stratification by cognitive impairment status showed the cumulative incidence of dementia was 17.4% (95% CI, 16.8-17.9) for none or minimal, 27.3% (95% CI, 26.9-27.8) for mild, and 35.3% (95% CI, 34.7-36.0) for moderate to severe.

CONCLUSIONS AND IMPLICATIONS

The cohort of people entering LTCFs with cognitive impairment had a high incidence of dementia diagnosis within 1.5 years after entry. Routine cognitive impairment assessments can inform dementia screening strategies by identifying individuals at higher risk of dementia.

Harnessing Electronic Health Records and Artificial Intelligence for Enhanced Cardiovascular Risk Prediction: A Comprehensive Review

Electronic health records (EHR) have revolutionized cardiovascular disease (CVD) research by enabling comprehensive, large-scale, and dynamic data collection. Integrating EHR data with advanced analytical methods, including artificial intelligence (AI), transforms CVD risk prediction and management methodologies. This review examines the advancements and challenges of using EHR in developing CVD prediction models, covering traditional and AI-based approaches. While EHR-based CVD risk prediction has greatly improved, moving from models that integrate real-world data on medication use and imaging, challenges persist regarding data quality, standardization across health care systems, and geographic variability. The complexity of EHR data requires sophisticated computational methods and multidisciplinary approaches for effective CVD risk modeling. AI's deep learning enhances prediction performance but faces limitations in interpretability and the need for validation and recalibration for diverse populations. The future of CVD risk prediction and management increasingly depends on using EHR and AI technologies effectively. Addressing data quality issues and overcoming limitations from retrospective data analysis are critical for improving the reliability and applicability of risk prediction models. Integrating multidimensional data, including environmental, lifestyle, social, and genomic factors, could significantly enhance risk assessment. These models require continuous validation and recalibration to ensure their adaptability to diverse populations and evolving health care environments, providing reassurance about their reliability.

Quality and Safety Indicators for Care Transitions by Older Adults: A Scoping Review

OBJECTIVE

To identify quality and safety indicators routinely used to monitor, evaluate, and improve care transitions for older adults globally.

DESIGN

A scoping literature review.

SETTING AND PARTICIPANTS

This review identified indicators used internationally to monitor and evaluate the quality and safety of care transitions by older adults. Care transitions were defined as the transfer of health care at least once between care settings.

METHODS

A search of academic and gray literature identified indicators that were publicly available, used routinely at the population level, and reported on since 2012. Indicators were summarized by care domain (ie, hospitalization, consumer experience, access/waiting times, communication, follow-up, and medication-related), type (structure, process, outcome), quality dimension (patient centeredness, timeliness, effectiveness, efficiency, safety, and equity), data collection approach, reporting strategies, and care settings involved.

RESULTS

The review identified 361 quality indicators from 89 programs across 12 countries. Care domains included hospitalization (n = 112; 31.0%), consumer experience (n = 82; 22.7%), access/waiting times (n = 63; 17.5%), communication (n = 40; 11.1%), follow-up (n = 40; 11.1%), and medication-related (n = 24; 6.6%). Indicators measured outcomes (n = 227; 62.9%) or processes (n = 134; 37.1%) and represented the dimensions of patient centeredness (n = 155, 42.9%), timeliness (n = 91; 25.2%), and effectiveness (n = 87; 24.1%), efficiency (n = 18; 5.0%) and safety (n = 10; 2.8%). Most indicators were constructed from survey (n = 160; 44.3%) or administrative data (n = 138; 38.2%); 69% (n = 249) were publicly reported and 80% (n = 287) measured transitions related to acute settings.

CONCLUSIONS AND IMPLICATIONS

Eighty-nine international programs routinely monitor the quality and safety of care transitions, and focus on the domains of hospitalization, access and waiting times, and communication. Considering the vulnerability of older adults as they transition across settings and providers, it is important to ensure holistic measurement of the quality of these care transitions to identify sub-optimal transitions, inform quality improvement, and ultimately improve outcomes for older adults.

"No Abnormality Detected": A Mixed-Methods Examination of Emergency Department Coding Practices for People in Suicidal Crisis

BACKGROUND

Accurate identification of suicidal crisis presentations to emergency departments (EDs) can lead to timely mental health support, improve patient experience, and support evaluations of suicide prevention initiatives. Poor coding practices within EDs are preventing appropriate patient care. Aims of the study are (1) examine the current suicide-related coding practices, (2) identify the factors that contribute to staff decision-making and patients receiving the incorrect code or no code.

METHOD

A mixed-methods study was conducted. Quantitative data were collated from six EDs across Merseyside and Cheshire, United Kingdom from 2019 to 2021. Attendances were analyzed if they had a presenting complaint, chief complaint, or primary diagnosis code related to suicidal crisis, suicidal ideation, self-harm or suicide attempt. Semi-structured interviews were conducted with staff holding various ED positions (n = 23).

RESULTS

A total of 15,411 suicidal crisis and self-harm presentations were analyzed. Of these, 21.8% were coded as 'depressive disorder' and 3.8% as 'anxiety disorder'. Absence of an appropriate suicidal crisis code resulted in staff coding presentations as 'no abnormality detected' (23.6%) or leaving the code blank (18.4%). The use of other physical injury codes such as 'wound forearm', 'head injury' were common. Qualitative analyses elucidated potential causes of inappropriate coding, such as resource constraints and problems with the recording process.

CONCLUSION

People attending EDs in suicidal crisis were not given a code that represented the chief presentation. Improved ED coding practices related to suicidal crisis could result in considerable benefits for patients and more effective targeting of resources and interventions.

Monitoring low-value care in medical patients from Swiss university hospitals using a Findable, Accessible, Interoperable, Reusable (FAIR) national data stream and patient and public involvement: LUCID study protocol

INTRODUCTION

Healthcare practices providing minimal or no benefit to recipients have been estimated to represent 20% of healthcare costs. However, defining, measuring and monitoring low-value care (LVC) and its downstream consequences remain a major challenge. The purpose of the National Data Stream (LUCID NDS) is to identify and monitor LVC in medical inpatients using routinely collected hospital data.

METHODS AND ANALYSIS

This protocol describes a multistep approach to the identification and surveillance of LVC: (1) creating an NDS based on Findable, Accessible, Interoperable, Reusable (FAIR) principles using routinely collected hospital data from medical inpatients who signed a general consent for data reuse from 2014 onwards; (2) selecting recommendations applicable to medical inpatients using data from LUCID NDS to develop a comprehensive and robust set of LVC indicators; (3) establishing expert consensus on the most relevant and actionable recommendations to prevent LVC; (4) applying the Strength of Recommendation Taxonomy methodology to assess the level of evidence of recommendations; (5) involving patients and the public at various stages of LUCID NDS; and (6) designing monitoring rules within the LUCID NDS and validating quality measures.

ETHICS AND DISSEMINATION

The ethics committees of all five participating university hospitals (Basel, Bern, Geneva, Lausanne and Zurich) approved LUCID NDS as a national registry on quality of care. We will disseminate our findings in peer-reviewed journals, at professional conferences, and through short reports sent to participating entities and stakeholders; moreover, lay summaries are provided for patients and the broader public on our webpage (www.LUCID-nds.ch).

Development of a National Colorectal Cancer Screening Research Agenda: An Initiative of the Canadian Screening for Colorectal Cancer Research Network (CanSCCRN)

The Canadian Screening for Colorectal Cancer Research Network (CanSCCRN) recently set out to develop a national CRC screening research agenda and identify priority research areas. The specific objectives were to (1) identify evidence gaps relevant to CRC screening and the barriers and facilitators to evidence generation and uptake by CRC screening programs, (2) establish high-priority collaborative research ideas to inform best CRC screening practices, and (3) identify one to two research topics for grant development and submission within 12 to 18 months. Three focus groups were conducted with network members and relevant parties (n = 15) to identify evidence gaps, barriers, and facilitators to evidence generation and uptake. Three workshops were subsequently held to discuss focus group findings and develop an action plan for research. An electronic survey was used to prioritize the evidence gaps to be addressed. Overall, five categories of barriers and six categories of facilitators to evidence uptake and generation were identified, as well as 23 evidence gaps to be addressed. Screening participation, post-polypectomy surveillance, and screening age range were identified as research priority research areas. Adequate resourcing and infrastructure, as well as partnerships with knowledge end users, are integral to addressing these research areas and advancing CRC screening programs in Canada and beyond.

Validation of the rates of adverse event incidence in administrative healthcare data through patient chart review: A scoping review protocol

Background

Patient safety is a key issue for health systems and a growing global public health challenge. Administrative healthcare data provide a coded summary of a patient and their encounter with the healthcare system. These aggregated datasets are often used to inform research and decisions relating to health service planning and therefore it is vital that they are accurate and reliable. Given the reported inaccuracy of these datasets for detecting and recording adverse events, there have been calls for validation studies to explore their reliability and investigate further their potential to inform research and health policy. Researchers have since carried out validation studies on the rates of adverse events in administrative data through chart reviews therefore, it seems appropriate to identify and chart the evidence and results of these studies within a scoping review.

Methods

The scoping review will be conducted in accordance with the Joanna Briggs Institute (JBI) methodology for scoping reviews. A search of databases such as PubMed, CINAHL, ScienceDirect and Scopus will be conducted in addition to a search of the reference lists of sourced publications and a search for grey literature. Following this, Covidence will be used to screen the sourced publications and subsequently extract data from the included sources. A numerical summary of the literature will be presented in addition to a charting based on the qualitative content analysis of the studies included.

Conclusions

This protocol provides the structure for the conduct of a review to identify and chart the evidence on validation studies on rates of adverse events in administrative healthcare data. This review will aim to identify research gaps, chart the evidence of and highlight any flaws within administrative datasets to improve extraction and coding practices and enable researchers and policy makers to use these data to their full potential.

Variation in Case Exposure During Internal Medicine Residency

Importance

Variation in residency case exposure affects resident learning and readiness for future practice. Accurate reporting of case exposure for internal medicine (IM) residents is challenging because feasible and reliable methods for linking patient care to residents are lacking.

Objective

To develop an integrated education-clinical database to characterize and measure case exposure variability among IM residents.

Design, Setting, and Participants

In this cohort study, an integrated educational-clinical database was developed by linking patients admitted during overnight IM in-hospital call shifts at 5 teaching hospitals to senior on-call residents. The senior resident, who directly cares for all overnight IM admissions, was linked to their patients by the admission date, time, and hospital. The database included IM residents enrolled between July 1, 2010, and December 31, 2019, in 1 Canadian IM residency. Analysis occurred between August 1, 2023, and June 30, 2024.

Main Outcomes and Measures

Case exposure was defined by patient demographic characteristics, discharge diagnoses, volumes, acuity (eg, critical care transfer), medical complexity (eg, Charlson Comorbidity Index), and social determinants of health (eg, from long-term care). Residents were grouped into quartiles for each exposure measure, and the top and bottom quartiles were compared using standardized mean difference (SMD). Variation between hospitals was evaluated by calculating the SMD between the hospitals with the highest and lowest proportions for each measure. Variation over time was assessed using linear and logistic regression.

Results

The integrated educational-clinical database included 143 632 admissions (median [IQR] age, 71 [55-83] years; 71 340 [49.7%] female) linked to 793 residents (median [IQR] admissions per shift, 8 [6-12]). At the resident level, there was substantial variation in case exposure for demographic characteristics, diagnoses, volumes, acuity, complexity, and social determinants. For example, residents in the highest quartile had nearly 4 times more admissions requiring critical care transfer compared with the lowest quartile (3071 of 30 228 [10.2%] vs 684 of 25 578 [2.7%]; SMD, 0.31). Hospital-level variation was also significant, particularly in patient volumes (busier hospital vs less busy hospital: median [IQR] admissions per shift, 10 [8-12] vs 7 [5-9]; SMD, 0.96). Over time, residents saw more median (IQR) admissions per shift (2010 vs 2019: 7.6 [6.6-8.4] vs 9.0 [7.6-10.0]; P = .04) and more complex patients (2010 vs 2019: Charlson Comorbidity Index ≥2, 3851 of 13 762 [28.0%] vs 2862 of 8188 [35.0%]; P = .03), while working similar shifts per year (median [IQR], 11 [8-14]).

Conclusions

In this cohort study of IM residents in a Canadian residency program, significant variation in case exposure was found between residents, across sites, and over time.

From prediction to practice: mitigating bias and data shift in machine-learning models for chemotherapy-induced organ dysfunction across unseen cancers

Objectives

Routine monitoring of renal and hepatic function during chemotherapy ensures that treatment-related organ damage has not occurred and clearance of subsequent treatment is not hindered; however, frequency and timing are not optimal. Model bias and data heterogeneity concerns have hampered the ability of machine learning (ML) to be deployed into clinical practice. This study aims to develop models that could support individualised decisions on the timing of renal and hepatic monitoring while exploring the effect of data shift on model performance.

Methods and analysis

We used retrospective data from three UK hospitals to develop and validate ML models predicting unacceptable rises in creatinine/bilirubin post cycle 3 for patients undergoing treatment for the following cancers: breast, colorectal, lung, ovarian and diffuse large B-cell lymphoma.

Results

We extracted 3614 patients with no missing blood test data across cycles 1-6 of chemotherapy treatment. We improved on previous work by including predictions post cycle 3. Optimised for sensitivity, we achieve F2 scores of 0.7773 (bilirubin) and 0.6893 (creatinine) on unseen data. Performance is consistent on tumour types unseen during training (F2 bilirubin: 0.7423, F2 creatinine: 0.6820).

Conclusion

Our technique highlights the effectiveness of ML in clinical settings, demonstrating the potential to improve the delivery of care. Notably, our ML models can generalise to unseen tumour types. We propose gold-standard bias mitigation steps for ML models: evaluation on multisite data, thorough patient population analysis, and both formalised bias measures and model performance comparisons on patient subgroups. We demonstrate that data aggregation techniques have unintended consequences on model bias.

A scoping review of the methodological approaches used in retrospective chart reviews to validate adverse event rates in administrative data

Patient safety is a key quality issue for health systems. Healthcare acquired adverse events (AEs) compromise safety and quality; therefore, their reporting and monitoring is a patient safety priority. Although administrative datasets are potentially efficient tools for monitoring rates of AEs, concerns remain over the accuracy of their data. Chart review validation studies are required to explore the potential of administrative data to inform research and health policy. This review aims to present an overview of the methodological approaches and strategies used to validate rates of AEs in administrative data through chart review. This review was conducted in line with the Joanna Briggs Institute methodological framework for scoping reviews. Through database searches, 1054 sources were identified, imported into Covidence, and screened against the inclusion criteria. Articles that validated rates of AEs in administrative data through chart review were included. Data were extracted, exported to Microsoft Excel, arranged into a charting table, and presented in a tabular and descriptive format. Fifty-six studies were included. Most sources reported on surgical AEs; however, other medical specialties were also explored. Chart reviews were used in all studies; however, few agreed on terminology for the study design. Various methodological approaches and sampling strategies were used. Some studies used the Global Trigger Tool, a two-stage chart review method, whilst others used alternative single-, two-stage, or unclear approaches. The sources used samples of flagged charts (n = 24), flagged and random charts (n = 11), and random charts (n = 21). Most studies reported poor or moderate accuracy of AE rates. Some studies reported good accuracy of AE recording which highlights the potential of using administrative data for research purposes. This review highlights the potential for administrative data to provide information on AE rates and improve patient safety and healthcare quality. Nonetheless, further work is warranted to ensure that administrative data are accurate. The variation of methodological approaches taken, and sampling techniques used demonstrate a lack of consensus on best practice; therefore, further clarity and consensus are necessary to develop a more systematic approach to chart reviewing.

Applications of Machine Learning (ML) and Mathematical Modeling (MM) in Healthcare with Special Focus on Cancer Prognosis and Anticancer Therapy: Current Status and Challenges

The use of data-driven high-throughput analytical techniques, which has given rise to computational oncology, is undisputed. The widespread use of machine learning (ML) and mathematical modeling (MM)-based techniques is widely acknowledged. These two approaches have fueled the advancement in cancer research and eventually led to the uptake of telemedicine in cancer care. For diagnostic, prognostic, and treatment purposes concerning different types of cancer research, vast databases of varied information with manifold dimensions are required, and indeed, all this information can only be managed by an automated system developed utilizing ML and MM. In addition, MM is being used to probe the relationship between the pharmacokinetics and pharmacodynamics (PK/PD interactions) of anti-cancer substances to improve cancer treatment, and also to refine the quality of existing treatment models by being incorporated at all steps of research and development related to cancer and in routine patient care. This review will serve as a consolidation of the advancement and benefits of ML and MM techniques with a special focus on the area of cancer prognosis and anticancer therapy, leading to the identification of challenges (data quantity, ethical consideration, and data privacy) which are yet to be fully addressed in current studies.

Economic Efficiency of an Open-Source National Medical Lab Software in Canada

Although the Canada federal government has invested over $3.1 billion developing health information technology (HIT), all 10 provinces still have their own separate HIT systems, which are non-interoperable, expensive, and inconsistent. After first reviewing how these systems operate, this paper analyzes the costs and savings of integrating the common billing, lab results, and diagnostic imaging (BLD) functions of these separate systems using free and open-source software and proposes a system for this, HermesAPI. Currently, 8 provincial governments representing over 95% of Canada's population allow private companies to create their own electronic medical records (EMR) system and integrate with provincial BLD systems. This study found the cost to develop and maintain HermesAPI would be between CAD$610,000 to CAD$740,000, but would prevent CAD$120,000 per company per province in development costs for a total savings of $6.4 million. HermesAPI would lower barriers to entry for the HIT industry to increase competition, improve the quality of HIT products, and ultimately patient care. The proposed open-source approach of the HermesAPI is one option towards building a more interoperable, less expensive, and more consistent HIT system for Canada.

Real word challenges in integrating electronic medical record and administrative health data for regional quality improvement in diabetes: a retrospective cross-sectional analysis

BACKGROUND

Linked electronic medical records and administrative data have the potential to support a learning health system and data-driven quality improvement. However, data completeness and accuracy must first be assessed before their application. We evaluated the processes, feasibility, and limitations of linking electronic medical records and administrative data for the purpose of quality improvement within five specialist diabetes clinics in Edmonton, Alberta, a province known for its robust health data infrastructure.

METHODS

We conducted a retrospective cross-sectional analysis using electronic medical record and administrative data for individuals ≥ 18 years attending the clinics between March 2017 and December 2018. Descriptive statistics were produced for demographics, service use, diabetes type, and standard diabetes benchmarks. The systematic and iterative process of obtaining results is described.

RESULTS

The process of integrating electronic medical record with administrative data for quality improvement was found to be non-linear and iterative and involved four phases: project planning, information generating, limitations analysis, and action. After limitations analysis, questions were grouped into those that were answerable with confidence, answerable with limitations, and not answerable with available data. Factors contributing to data limitations included inaccurate data entry, coding, collation, migration and synthesis, changes in laboratory reporting, and information not captured in existing databases.

CONCLUSION

Electronic medical records and administrative databases can be powerful tools to establish clinical practice patterns, inform data-driven quality improvement at a regional level, and support a learning health system. However, there are substantial data limitations that must be addressed before these sources can be reliably leveraged.

External Validation of a Prediction Model for Falls in Older People Based on Electronic Health Records in Primary Care

OBJECTIVE

Early identification of older people at risk of falling is the cornerstone of fall prevention. Many fall prediction tools exist but their external validity is lacking. External validation is a prerequisite before application in clinical practice. Models developed with electronic health record (EHR) data are especially challenging because of the uncontrolled nature of routinely collected data. We aimed to externally validate our previously developed and published prediction model for falls, using a large cohort of community-dwelling older people derived from primary care EHR data.

DESIGN

Retrospective analysis of a prospective cohort drawn from EHR data.

SETTING AND PARTICIPANTS

Pseudonymized EHR data were collected from individuals aged ≥65 years, who were enlisted in any of the participating 59 general practices between 2015 and 2020 in the Netherlands.

METHODS

Ten predictors were defined and obtained using the same methods as in the development study. The outcome was 1-year fall and was obtained from free text. Both reproducibility and transportability were evaluated. Model performance was assessed in terms of discrimination using the area under the receiver operating characteristic curve (ROC-AUC), and in terms of calibration, using calibration-in-the-large, calibration slope and calibration plots.

RESULTS

Among 39,342 older people, 5124 (13.4%) fell in the 1-year follow-up. The characteristics of the validation and the development cohorts were similar. ROC-AUCs of the validation and development cohort were 0.690 and 0.705, respectively. Calibration-in-the-large and calibration slope were 0.012 and 0.878, respectively. Calibration plots revealed overprediction for high-risk groups in a small number of individuals.

CONCLUSIONS AND IMPLICATIONS

Our previously developed prediction model for falls demonstrated good external validity by reproducing its predictive performance in the validation cohort. The implementation of this model in the primary care setting could be considered after impact assessment.