Assessing the accuracy of administrative data in health information systems

Recent Trends in Diabetic and Nondiabetic Neuropathies: A Retrospective Hospital-based Nationwide Cohort Study

OBJECTIVE

The aim of this study was to evaluate the trends in the incidence of diabetic neuropathy (DN) and nondiabetic neuropathy (non-DN) in a hospital-based cohort between 2010 and 2019 in Romania.

METHODS

We retrospectively analyzed cases with a primary or secondary discharge International Classification of Diseases, Tenth Revision, diagnosis code of neuropathy reported throughout Romania.

RESULTS

A total of 1 725 729 hospitalizations with a diagnosis of neuropathy (DN, 769 324 - 44.6%, and non-DN, 956 405- 55.4%) were identified. Women accounted for more DN cases (40 0 936- 52.1%), and men accounted for more non-DN cases (63 7 968- 61.0%). The incidence rate showed an increasing trend during the index period, by a mean rate of 4.3%/year for non-DN and 1.4%/year for DN. Type 2 diabetes was responsible for the overall increase in the incidence rate of DN, whereas in type 1 diabetes, the incidence rate decreased; in both types of diabetes, diabetic polyneuropathy was predominant, whereas autonomic neuropathy had an incidence rate of 10 to 20 times lower than polyneuropathy. The incidence rates of non-DNs increased mainly due to inflammatory polyneuropathies (+3.8%) and uremic neuropathy (+10.3%).

CONCLUSION

Using a nationally representative database of hospital-admitted cases, we found that the incidence rates of DN and non-DN increased from 2010 to 2019. The main contributors were type 2 diabetes, inflammatory polyneuropathy, and uremic neuropathy.

Interprofessional Coproduction of Diagnosis with Medical and Pharmacy Students: An Interactive Case-Based Workshop

Introduction

The Institute of Medicine and national competencies emphasize the importance of interprofessional education to reduce diagnostic error. Clinical pharmacists are increasingly integrated into clinical teams and participate in the diagnostic process. However, few educational resources explicitly teach medical and pharmacy students to collaborate on the production of diagnoses.

Methods

We implemented a 2-hour, online, case-based workshop with 154 second-year medical students and third-year pharmacy students. After brief didactics on the diagnostic process and scope of practice of pharmacists, small groups of eight to 12 medical and pharmacy students with faculty facilitators worked through a case unfolding in two aliquots. Students were provided different but complementary information authentic to their profession. They had to communicate with each other to develop an appropriate differential diagnosis. Students then reflected on how communicating with the other profession impacted their diagnostic reasoning. Comments were coded and counted.

Results

The majority (99%) of students identified their data gathering and differential diagnoses were impacted by working through the case together. More pharmacy students commented on how medical students broadened their differential diagnosis (71%) and added information (72%), contextualizing information, such as past history, medication indications, and physical exam data. More medical students commented on how pharmacy students helped justify (54%) and clarify (22%) the differential diagnosis, often connecting the underlying mechanism of medications with clinical findings.

Discussion

This interactive case-based workshop was effective in teaching medical and pharmacy students to collaborate in the coproduction of diagnosis. It is feasible with minimal resources.

Consistency Between Administrative Health Records and Self-Reported Health Status and Health Care Use Among Indigenous Wayuu Health Insurance Enrollees: La Guajira, Colombia

The objective of the study was to assess the consistency between self-reported demographic characteristics, health conditions, and healthcare use, and administrative healthcare records, in a sample of enrollees of an Indigenous health organization in Colombia. We conducted a phone survey of a random sample of 2113 enrollees September-2020/February-2021. Administrative health records were obtained for the sample. Using ICD-10 diagnostic codes, we identified individuals who had healthcare visits for diabetes, hypertension, and/or pregnancy. Using unique identifiers, we linked their survey data to the administrative dataset. Agreement percentages and Cohen's Kappa coefficients were calculated. Logistic regressions were performed for each health condition/state. Results showed high degree of agreement between data sources for sex and age, similar rates for diabetes and hypertension, 10% variation for pregnancy. Kappa statistics were in the moderate range. Age was significantly associated with agreement between data sources. Sex, language, and self-rated health were significant for diabetes. This is the first study with data from an Indigenous population assessing the consistency between self-reported data and administrative health records. Survey and administrative data produced similar results, suggesting that Anas Wauu can be confident in using their data for planning and research purposes, as part of the movement toward data sovereignty.

Comparing broad and narrow phenotype algorithms: differences in performance characteristics and immortal time incurred

Introduction: When developing phenotype algorithms for observational research, there is usually a trade-off between definitions that are sensitive or specific. The objective of this study was to estimate the performance characteristics of phenotype algorithms designed for increasing specificity and to estimate the immortal time associated with each algorithm. Materials and methods: We examined algorithms for 11 chronic health conditions. The analyses were from data from five databases. For each health condition, we created five algorithms to examine performance (sensitivity and positive predictive value (PPV)) differences: one broad algorithm using a single code for the health condition and four narrow algorithms where a second diagnosis code was required 1-30 days, 1-90 days, 1-365 days, or 1- all days in a subject's continuous observation period after the first code. We also examined the proportion of immortal time relative to time-at-risk (TAR) for four outcomes. The TAR's were: 0-30 days after the first condition occurrence (the index date), 0-90 days post-index, 0-365 days post-index, and 0-1,095 days post-index. Performance of algorithms for chronic health conditions was estimated using PheValuator (V2.1.4) from the OHDSI toolstack. Immortal time was calculated as the time from the index date until the first of the following: 1) the outcome; 2) the end of the outcome TAR; 3) the occurrence of the second code for the chronic health condition. Results: In the first analysis, the narrow phenotype algorithms, i.e., those requiring a second condition code, produced higher estimates for PPV and lower estimates for sensitivity compared to the single code algorithm. In all conditions, increasing the time to the required second code increased the sensitivity of the algorithm. In the second analysis, the amount of immortal time increased as the window used to identify the second diagnosis code increased. The proportion of TAR that was immortal was highest in the 30 days TAR analyses compared to the 1,095 days TAR analyses. Conclusion: Attempting to increase the specificity of a health condition algorithm by adding a second code is a potentially valid approach to increase specificity, albeit at the cost of incurring immortal time.

Identifying patients with opioid use disorder using International Classification of Diseases (ICD) codes: Challenges and opportunities

BACKGROUND AND AIMS

International Classification of Diseases (ICD) diagnosis codes are often used in research to identify patients with opioid use disorder (OUD), but their accuracy for this purpose is not fully evaluated. This study describes application of ICD-10 diagnosis codes for opioid use, dependence and abuse from an electronic health record (EHR) data extraction using data from the clinics' OUD patient registries and clinician/staff EHR entries.

DESIGN

Cross-sectional observational study.

SETTING

Four rural primary care clinics in Washington and Idaho, USA.

PARTICIPANTS

307 patients.

MEASUREMENTS

This study used three data sources from each clinic: (1) a limited dataset extracted from the EHR, (2) a clinic-based registry of patients with OUD and (3) the clinician/staff interface of the EHR (e.g. progress notes, problem list). Data source one included records with six commonly applied ICD-10 codes for opioid use, dependence and abuse: F11.10 (opioid abuse, uncomplicated), F11.20 (opioid dependence, uncomplicated), F11.21 (opioid dependence, in remission), F11.23 (opioid dependence with withdrawal), F11.90 (opioid use, unspecified, uncomplicated) and F11.99 (opioid use, unspecified with unspecified opioid-induced disorder). Care coordinators used data sources two and three to categorize each patient identified in data source one: (1) confirmed OUD diagnosis, (2) may have OUD but no confirmed OUD diagnosis, (3) chronic pain with no evidence of OUD and (4) no evidence for OUD or chronic pain.

FINDINGS

F11.10, F11.21 and F11.99 were applied most frequently to patients who had clinical diagnoses of OUD (64%, 89% and 79%, respectively). F11.20, F11.23 and F11.90 were applied to patients who had a diagnostic mix of OUD and chronic pain without OUD. The four clinics applied codes inconsistently.

CONCLUSIONS

Lack of uniform application of ICD diagnosis codes make it challenging to use diagnosis code data from EHR to identify a research population of persons with opioid use disorder.

Using a data-driven approach for the development and evaluation of phenotype algorithms for systemic lupus erythematosus

BACKGROUND

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease of unknown origin. The objective of this research was to develop phenotype algorithms for SLE suitable for use in epidemiological studies using empirical evidence from observational databases.

METHODS

We used a process for empirically determining and evaluating phenotype algorithms for health conditions to be analyzed in observational research. The process started with a literature search to discover prior algorithms used for SLE. We then used a set of Observational Health Data Sciences and Informatics (OHDSI) open-source tools to refine and validate the algorithms. These included tools to discover codes for SLE that may have been missed in prior studies and to determine possible low specificity and index date misclassification in algorithms for correction.

RESULTS

We developed four algorithms using our process: two algorithms for prevalent SLE and two for incident SLE. The algorithms for both incident and prevalent cases are comprised of a more specific version and a more sensitive version. Each of the algorithms corrects for possible index date misclassification. After validation, we found the highest positive predictive value estimate for the prevalent, specific algorithm (89%). The highest sensitivity estimate was found for the sensitive, prevalent algorithm (77%).

CONCLUSION

We developed phenotype algorithms for SLE using a data-driven approach. The four final algorithms may be used directly in observational studies. The validation of these algorithms provides researchers an added measure of confidence that the algorithms are selecting subjects correctly and allows for the application of quantitative bias analysis.

Current Challenges and Future Direction in Surveillance for Hepatocellular Carcinoma in Patients with Nonalcoholic Fatty Liver Disease

The burden for hepatocellular carcinoma (HCC) attributed to nonalcoholic fatty liver disease (NAFLD) continues to grow in parallel with rising global trends in obesity. The risk of HCC is elevated among patients with NAFLD-related cirrhosis to a level that justifies surveillance based on cost-effectiveness argument. The quality of current evidence for HCC surveillance in all patients with chronic liver disease is poor, and even lower in those with NAFLD. For a lack of more precise risk-stratification tools, current approaches to defining a target population in noncirrhotic NAFLD are limited to noninvasive tests for liver fibrosis, as a proxy for liver-related morbidity and mortality. Beyond etiology and severity of liver disease, traditional and metabolic risk factors, such as diabetes mellitus, older age, male gender and tobacco smoking, are not enough for HCC risk stratification for surveillance efficacy and effectiveness in NAFLD. There is an association between molecular and genetic factors and HCC risk in NAFLD, and risk models integrating both clinical and genetic factors will be key to personalizing HCC risk. In this review, we discuss concerns regarding defining a target population, surveillance test accuracy, surveillance underuse, and other cost-effective considerations for HCC surveillance in individuals with NAFLD.

Inconsistent Surgical Implant Documentation: A Case Study in Total Knee and Hip Arthroplasty

Value-based care initiatives require accurate quantification of resource utilization. This study explores hospital resource documentation performance for total knee and hip arthroplasty (TKA, THA) implants and how this may differ between hospitals. This retrospective study utilized the Premier discharge database, years 2006 to 2020. TKA/THA cases were categorized into 5 tiers based upon the completeness of implant component documentation: Platinum, Gold, Silver, Bronze, Poor. Correlation between TKA and THA documentation performance (per-hospital percentage of Platinum cases) was assessed. Logistic regression analyses measured the association between hospital characteristics (region, teaching status, bed size, urban/rural) and satisfactory documentation. TKA/THA implant documentation performance was compared to documentation for endovascular stent procedures. Individual hospitals tended to have very complete (Platinum) or very incomplete (Poor) documentation for both TKA and THA. TKA and THA documentation performance were correlated (correlation coefficient = .70). Teaching hospitals were less likely to have satisfactory documentation for both TKA (P = .002) and THA (P = .029). Documentation for endovascular stent procedures was superior compared to TKA/THA. Hospitals' TKA and THA-related implant documentation performance is generally either very proficient or very poor, in contrast with often well-documented endovascular stent procedures. Hospital characteristics, other than teaching status, do not appear to impact TKA/THA documentation completeness.

Estimated costs of preoperative evaluation of postmenopausal hysterectomy for prolapse at a safety-net hospital: an observational descriptive study

BACKGROUND

In practice, preoperative evaluation prior to hysterectomy varies. Unnecessary preoperative evaluation may add cost and risk with little benefit to the patient.

OBJECTIVE

This study aimed to describe practice patterns and the associated costs related to preoperative evaluations before hysterectomy for prolapse at a safety-net hospital.

STUDY DESIGN

This was a retrospective cohort study of postmenopausal women who underwent a hysterectomy for prolapse. Nonfacility-associated cost data were obtained from the Centers for Medicare Services. The biopsy cost was estimated to be $172.55 and $125.23 for ultrasounds.

RESULTS

A total of 505 postmenopausal cases were identified. Of those, 155 (31%) underwent a preoperative biopsy, 305 (60%) had an ultrasound, and 124 (25%) had both. Of those, 72.9% had an indication for a biopsy. A total of 64 biopsies and 216 ultrasounds lacked clear indication. Of those, 56 biopsies were performed for bleeding in cases with an endometrial thickness of <4 mm. The total cost of nonvalue-added testing was $42,576.

CONCLUSION

Adherence to a strict preoperative algorithm would have saved $38,092 over the study period, although 0.50% of these biopsies would potentially have detected endometrial cancer preoperatively. These results underscore the value of clinical algorithms at teaching institutions.

An analysis of government-sponsored health insurance enrolment and claims data from Meghalaya: Insights into the provision of health care in North East India

Introduction

The Megha Health Insurance Scheme (MHIS) was launched in 2013 in the North-East Indian state of Meghalaya to reduce household out-of-pocket expenditure on health and provide access to high-quality essential healthcare. Despite substantial expansion of the MHIS since the scheme’s inception, there is a lack of comprehensive documentation and evaluation of the scheme’s performance against its Universal Health Care (UHC) objectives.

Methods

We analysed six years of enrolment and claims data (2013–2018) covering three phases of the scheme to understand the pattern of enrolment, utilisation and care provision under the MHIS during this period. De-identified data files included information on age, sex, district of residence, the district of provider hospital, type of hospital, date of admission, status at discharge, claimed category of care, package codes, and amount claimed. Descriptive statistics were generated to investigate key trends in enrolment, service utilisation, and Government health spending under the MHIS.

Results

Approximately 55% of the eligible population are currently enrolled in MHIS. Enrolment increased consistently from phase I through III and remained broadly stable across districts, gender, age group and occupation categories, with a small decline in males 19–60 years. Claims were disproportionately skewed towards private provision; 57% of all claims accrued to the 18 empanelled private hospitals and 39% to the 159 public sector facilities. The package ‘General Ward Unspecified’ was responsible for the highest volume of claims and highest financial dispensation across all three phases of the scheme. This likely indicates substantial administrative error and is potentially masking both true burden of disease and accurate financial provision for care under the MHIS. Anti-rabies injections for dog/cat bite contributed to 11% of total claims under MHIS III, and 1.6% of all claims under MHIS II. This warrants investigation to better understand the burden of animal bites on the Meghalayan population and inform the implementation of cost-effective strategies to reduce this burden.

Conclusions

This paper describes the first analysis of health insurance enrolment and claims data in the state of Meghalaya. The analysis has generated an important evidence base to inform future MHIS enrolment and care provision policies as the scheme expands to provide Universal Health Coverage to the state’s entire population.

A multicenter evaluation of computable phenotyping approaches for SARS-CoV-2 infection and COVID-19 hospitalizations

Diagnosis codes are used to study SARS-CoV2 infections and COVID-19 hospitalizations in administrative and electronic health record (EHR) data. Using EHR data (April 2020-March 2021) at the Yale-New Haven Health System and the three hospital systems of the Mayo Clinic, computable phenotype definitions based on ICD-10 diagnosis of COVID-19 (U07.1) were evaluated against positive SARS-CoV-2 PCR or antigen tests. We included 69,423 patients at Yale and 75,748 at Mayo Clinic with either a diagnosis code or a positive SARS-CoV-2 test. The precision and recall of a COVID-19 diagnosis for a positive test were 68.8% and 83.3%, respectively, at Yale, with higher precision (95%) and lower recall (63.5%) at Mayo Clinic, varying between 59.2% in Rochester to 97.3% in Arizona. For hospitalizations with a principal COVID-19 diagnosis, 94.8% at Yale and 80.5% at Mayo Clinic had an associated positive laboratory test, with secondary diagnosis of COVID-19 identifying additional patients. These patients had a twofold higher inhospital mortality than based on principal diagnosis. Standardization of coding practices is needed before the use of diagnosis codes in clinical research and epidemiological surveillance of COVID-19.

An Observational Study of the Association Between Exposure to Vacant Building Demolitions and Elevated Blood Lead Levels in Children Under Six in St Louis City

CONTEXT

St Louis City has been demolishing vacant buildings at an increasing rate. Demolition can cause lead dust spread, and childhood lead exposure can have negative effects on cognition, growth, and development. Previous studies show an association between exposure to multiple demolitions and elevated blood lead levels (EBLLs) in children, but St Louis City does not monitor the effects of demolitions on children's blood lead levels.

OBJECTIVES

The purpose of this study was to measure the association between exposure to demolitions and EBLLs in children younger than 6 years in St Louis City from 2017 to 2020.

DESIGN/SETTING/PARTICIPANTS

We analyzed blood lead testing data for children 0 to 72 months of age (n = 22 192) and proximity to demolitions. Exposure was the presence of demolitions within 400 ft of a child's address in the 33 days before their first lead test.

MAIN OUTCOME MEASURE

We used logistic regression to test the association between proximity to demolition and EBLLs (≥5 µg/dL).

RESULTS

The percentage of children living in proximity to 1 or more demolitions was slightly higher among those with EBLLs (n = 21; 1.3%) than among those without EBLLs (n = 250; 1.2%). However, after adjusting for age, sex, year home was built, season, neighborhood socioeconomic percentile, and neighborhood racial composition, the odds of EBLLs were not significantly different for children exposed to 1 or more demolitions (OR = 0.82; 95% CI, 0.5-1.25) compared with exposed to zero demolitions.

CONCLUSIONS

Although this study found no association between exposure to demolitions and EBLLs, results should be interpreted with caution, given numerous limitations. Given the consequences of childhood lead exposure, it is recommended that St Louis City conduct a similar analysis on demolitions conducted after 2020 using systematically collected demolition dates. Targeted testing or soil and air monitoring could also be informative.

Reporting transparency and completeness in trials: Paper 3 - trials conducted using administrative databases do not adequately report elements related to use of databases

OBJECTIVE

We evaluated reporting completeness and transparency in randomized controlled trials (RCTs) conducted using administrative data based on 2021 CONSORT Extension for Trials Conducted Using Cohorts and Routinely Collected Data (CONSORT-ROUTINE) criteria.

STUDY DESIGN AND SETTING

MEDLINE and the Cochrane Methodology Register were searched (2011 and 2018). Eligible RCTs used administrative databases for identifying eligible participants or collecting outcomes. We evaluated reporting based on CONSORT-ROUTINE, which modified eight items from CONSORT 2010 and added five new items.

RESULTS

Of 33 included trials (76% used administrative databases for outcomes, 3% for identifying participants, 21% both), most were conducted in the United States (55%), Canada (18%), or the United Kingdom (12%). Of eight items modified in the extension; six were adequately reported in a majority (>50%) of trials. For the CONSORT-ROUTINE modification portion of those items, three items were reported adequately in >50% of trials, two in <50%, two only applied to some trials, and one only had wording modifications and was not evaluated. For five new items, four that address use of routine data in trials were reported inadequately in most trials.

CONCLUSION

How administrative data are used in trials is often sub-optimally reported. CONSORT-ROUTINE uptake may improve reporting.

Development of a Semiautomated Search Tool to Identify Grading From Pathology Reports for Tumors of the CNS and Prostate Cancers

PURPOSE

This study demonstrates the functionality of semiautomated algorithms to classify cancer-specific grading from electronic pathology reports generated from military treatment facilities. Two Perl-based algorithms are validated to classify WHO grade for tumors of the CNS and Gleason grades for prostate cancer.

METHODS

Case-finding cohorts were developed using diagnostic codes and matched by unique identifiers to obtain pathology records generated in the Military Health System for active duty service members from 2013 to 2018. Perl-based algorithms were applied to classify document-based pathology reports to identify malignant CNS tumors and prostate cancer, followed by a hand-review process to determine accuracy of the algorithm classifications. Inter-rater reliability, sensitivity, specificity, positive predictive values (PPVs), and negative predictive values were computed following abstractor adjudication.

RESULTS

The high PPV for the Perl-based algorithms to classify CNS tumors (PPV > 98%) and prostate cancer (PPV > 99%) supports this approach to classify malignancies for cancer surveillance operations, mediated by a hand-reviewed semiautomated process to increase sensitivity by capturing ungraded cancers. Early detection was pronounced where 33.6% and 50.7% of malignant records retained a CNS WHO grade of II or a Gleason score of 6, respectively. Sensitivity metrics met criteria (> 75%) for brain (79.9%, 95% CI, 73.0 to 85.7) and prostate (96.7%, 95% CI, 94.9 to 98.0) cancers.

CONCLUSION

Semiautomated, document-based text classification using Perl coding successfully leveraged identification of WHO and Gleason grades to classify pathology records for CNS tumors and prostate cancer. The process is recommended for data quality initiatives to support cancer reporting functions, epidemiology, and research.

A Phenotyping Algorithm to Identify People With HIV in Electronic Health Record Data (HIV-Phen): Development and Evaluation Study

Background

Identification of people with HIV from electronic health record (EHR) data is an essential first step in the study of important HIV outcomes, such as risk assessment. This task has been historically performed via manual chart review, but the increased availability of large clinical data sets has led to the emergence of phenotyping algorithms to automate this process. Existing algorithms for identifying people with HIV rely on a combination of International Classification of Disease codes and laboratory tests or closely mimic clinical testing guidelines for HIV diagnosis. However, we found that existing algorithms in the literature missed a significant proportion of people with HIV in our data.

Objective

The aim of this study is to develop and evaluate HIV-Phen, an updated criteria-based HIV phenotyping algorithm.

Methods

We developed an algorithm using HIV-specific laboratory tests and medications and compared it with previously published algorithms in national and local data sets to identify cohorts of people with HIV. Cohort demographics were compared with those reported in the national and local surveillance data. Chart reviews were performed on a subsample of patients from the local database to calculate the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the algorithm.

Results

Our new algorithm identified substantially more people with HIV in both national (up to an 85.75% increase) and local (up to an 83.20% increase) EHR databases than the previously published algorithms. The demographic characteristics of people with HIV identified using our algorithm were similar to those reported in national and local HIV surveillance data. Our algorithm demonstrated improved sensitivity over existing algorithms (98% vs 56%-92%) while maintaining a similar overall accuracy (96% vs 80%-96%).

Conclusions

We developed and evaluated an updated criteria-based phenotyping algorithm for identifying people with HIV in EHR data that demonstrates improved sensitivity over existing algorithms.

Diagnostic error in the pediatric hospital: a narrative review

INTRODUCTION

Diagnostic error is a prevalent type of medical error that is associated with considerable patient harm and increased medical costs. The majority of literature guiding the current understanding of diagnostic error in the hospital setting is from adult studies. However, there is research to suggest this type of error is also prevalent in the pediatric specialty.

OBJECTIVES

The primary objective of this study was to define the current understanding of diagnostic error in the pediatric hospital through a structured literature review.

METHODS

We searched PubMed and identified studies focusing on three aspects of diagnostic error in pediatric hospitals: the incidence or prevalence, contributing factors, and related interventions. We used a tiered review, and a standardized electronic form to extract data from included articles.

RESULTS

Fifty-nine abstracts were screened and 23 full-text studies were included in the final review. Seventeen of the 23 studies focused on the incidence or prevalence, with only 3 studies investigating the utility of interventions. Most studies took place in an intensive care unit or emergency department with very few studies including only patients on the general wards. Overall, the prevalence of diagnostic error in pediatric hospitals varied greatly and depended on the measurement technique and specific hospital setting. Both healthcare system factors and individual cognitive factors were found to contribute to diagnostic error, with there being limited evidence to guide how best to mitigate the influence of these factors on the diagnostic process.

CONCLUSION

The general knowledge of diagnostic error in pediatric hospital settings is limited. Future work should incorporate structured frameworks to measure diagnostic errors and examine clinicians' diagnostic processes in real-time to help guide effective hospital-wide interventions.

Optimizing Identification of People Living with HIV from Electronic Medical Records: Computable Phenotype Development and Validation

BACKGROUND

Electronic health record (EHR)-based computable phenotype algorithms allow researchers to efficiently identify a large virtual cohort of Human Immunodeficiency Virus (HIV) patients. Built upon existing algorithms, we refined, improved, and validated an HIV phenotype algorithm using data from the OneFlorida Data Trust, a repository of linked claims data and EHRs from its clinical partners, which provide care to over 15 million patients across all 67 counties in Florida.

METHODS

Our computable phenotype examined information from multiple EHR domains, including clinical encounters with diagnoses, prescription medications, and laboratory tests. To identify an HIV case, the algorithm requires the patient to have at least one diagnostic code for HIV and meet one of the following criteria: have 1+ positive HIV laboratory, have been prescribed with HIV medications, or have 3+ visits with HIV diagnostic codes. The computable phenotype was validated against a subset of clinical notes.

RESULTS

Among the 15+ million patients from OneFlorida, we identified 61,313 patients with confirmed HIV diagnosis. Among them, 8.05% met all four inclusion criteria, 69.7% met the 3+ HIV encounters criteria in addition to having HIV diagnostic code, and 8.1% met all criteria except for having positive laboratories. Our algorithm achieved higher sensitivity (98.9%) and comparable specificity (97.6%) relative to existing algorithms (77-83% sensitivity, 86-100% specificity). The mean age of the sample was 42.7 years, 58% male, and about half were Black African American. Patients' average follow-up period (the time between the first and last encounter in the EHRs) was approximately 4.6 years. The median number of all encounters and HIV-related encounters were 79 and 21, respectively.

CONCLUSION

By leveraging EHR data from multiple clinical partners and domains, with a considerably diverse population, our algorithm allows more flexible criteria for identifying patients with incomplete laboratory test results and medication prescribing history compared with prior studies.

Accuracy of Computable Phenotyping Approaches for SARS-CoV-2 Infection and COVID-19 Hospitalizations from the Electronic Health Record

Objective:

Real-world data have been critical for rapid-knowledge generation throughout the COVID-19 pandemic. To ensure high-quality results are delivered to guide clinical decision making and the public health response, as well as characterize the response to interventions, it is essential to establish the accuracy of COVID-19 case definitions derived from administrative data to identify infections and hospitalizations.

Methods:

Electronic Health Record (EHR) data were obtained from the clinical data warehouse of the Yale New Haven Health System (Yale, primary site) and 3 hospital systems of the Mayo Clinic (validation site). Detailed characteristics on demographics, diagnoses, and laboratory results were obtained for all patients with either a positive SARS-CoV-2 PCR or antigen test or ICD-10 diagnosis of COVID-19 (U07.1) between April 1, 2020 and March 1, 2021. Various computable phenotype definitions were evaluated for their accuracy to identify SARS-CoV-2 infection and COVID-19 hospitalizations.

Results:

Of the 69,423 individuals with either a diagnosis code or a laboratory diagnosis of a SARS-CoV-2 infection at Yale, 61,023 had a principal or a secondary diagnosis code for COVID-19 and 50,355 had a positive SARS-CoV-2 test. Among those with a positive laboratory test, 38,506 (76.5%) and 3449 (6.8%) had a principal and secondary diagnosis code of COVID-19, respectively, while 8400 (16.7%) had no COVID-19 diagnosis. Moreover, of the 61,023 patients with a COVID-19 diagnosis code, 19,068 (31.2%) did not have a positive laboratory test for SARS-CoV-2 in the EHR. Of the 20 cases randomly sampled from this latter group for manual review, all had a COVID-19 diagnosis code related to asymptomatic testing with negative subsequent test results. The positive predictive value (precision) and sensitivity (recall) of a COVID-19 diagnosis in the medical record for a documented positive SARS-CoV-2 test were 68.8% and 83.3%, respectively. Among 5,109 patients who were hospitalized with a principal diagnosis of COVID-19, 4843 (94.8%) had a positive SARS-CoV-2 test within the 2 weeks preceding hospital admission or during hospitalization. In addition, 789 hospitalizations had a secondary diagnosis of COVID-19, of which 446 (56.5%) had a principal diagnosis consistent with severe clinical manifestation of COVID-19 (e.g., sepsis or respiratory failure). Compared with the cohort that had a principal diagnosis of COVID-19, those with a secondary diagnosis had a more than 2-fold higher in-hospital mortality rate (13.2% vs 28.0%, P<0.001). In the validation sample at Mayo Clinic, diagnosis codes more consistently identified SARS-CoV-2 infection (precision of 95%) but had lower recall (63.5%) with substantial variation across the 3 Mayo Clinic sites. Similar to Yale, diagnosis codes consistently identified COVID-19 hospitalizations at Mayo, with hospitalizations defined by secondary diagnosis code with 2-fold higher in-hospital mortality compared to those with a primary diagnosis of COVID-19.

Conclusions:

COVID-19 diagnosis codes misclassified the SARS-CoV-2 infection status of many people, with implications for clinical research and epidemiological surveillance. Moreover, the codes had different performance across two academic health systems and identified groups with different risks of mortality. Real-world data from the EHR can be used to in conjunction with diagnosis codes to improve the identification of people infected with SARS-CoV-2.

Validity of Incident Opioid Use Disorder (OUD) Diagnoses in Administrative Data: a Chart Verification Study

BACKGROUND

An important strategy to address the opioid overdose epidemic involves identifying people at elevated risk of overdose, particularly those with opioid use disorder (OUD). However, it is unclear to what degree OUD diagnoses in administrative data are inaccurate.

OBJECTIVE

To estimate the prevalence of inaccurate diagnoses of OUD among patients with incident OUD diagnoses.

SUBJECTS

A random sample of 90 patients with incident OUD diagnoses associated with an index in-person encounter between October 1, 2016, and June 1, 2018, in three Veterans Health Administration medical centers.

DESIGN

Direct chart review of all encounter notes, referrals, prescriptions, and laboratory values within a 120-day window before and after the index encounter. Using all available chart data, we determined whether the diagnosis of OUD was likely accurate, likely inaccurate, or of indeterminate accuracy. We then performed a bivariate analysis to assess demographic or clinical characteristics associated with likely inaccurate diagnoses.

KEY RESULTS

We identified 1337 veterans with incident OUD diagnoses. In the chart verification subsample, we assessed 26 (29%) OUD diagnoses as likely inaccurate; 20 due to systems error and 6 due to clinical error; additionally, 8 had insufficient information to determine accuracy. Veterans with likely inaccurate diagnoses were more likely to be younger and prescribed opioids for pain. Clinical settings associated with likely inaccurate diagnoses were non-mental health clinical settings, group visits, and non-patient care settings.

CONCLUSIONS

Our study identified significant levels of likely inaccurate OUD diagnoses among veterans with incident OUD diagnoses. The majority of these cases reflected readily addressable systems errors. The smaller proportion due to clinical errors and those with insufficient documentation may be addressed by increased training for clinicians. If these inaccuracies are prevalent throughout the VHA, they could complicate health services research and health systems responses.

A Review on the Application and Limitations of Administrative Health Care Data for the Study of Acute Kidney Injury Epidemiology and Outcomes in Children

Administrative health care databases contain valuable patient information generated by health care encounters. These "big data" repositories have been increasingly used in epidemiological health research internationally in recent years as they are easily accessible and cost-efficient and cover large populations for long periods. Despite these beneficial characteristics, it is also important to consider the limitations that administrative health research presents, such as issues related to data incompleteness and the limited sensitivity of the variables. These barriers potentially lead to unwanted biases and pose threats to the validity of the research being conducted. In this review, we discuss the effectiveness of health administrative data in understanding the epidemiology of and outcomes after acute kidney injury (AKI) among adults and children. In addition, we describe various validation studies of AKI diagnostic or procedural codes among adults and children. These studies reveal challenges of AKI research using administrative data and the lack of this type of research in children and other subpopulations. Additional pediatric-specific validation studies of administrative health data are needed to promote higher volume and increased validity of this type of research in pediatric AKI, to elucidate the large-scale epidemiology and patient and health systems impacts of AKI in children, and to devise and monitor programs to improve clinical outcomes and process of care.

Evaluation of the malaria reporting system supported by the District Health Information System 2 in Solomon Islands

Background

District Health Information Systems 2 (DHIS2) is used for supporting health information management in 67 countries, including Solomon Islands. However, there have been few published evaluations of the performance of DHIS2-enhanced disease reporting systems, in particular for monitoring infectious diseases such as malaria. The aim of this study was to evaluate DHIS2 supported malaria reporting in Solomon Islands and to develop recommendations for improving the system.

Methods

The evaluation was conducted in three administrative areas of Solomon Islands: Honoria City Council, and Malaita and Guadalcanal Provinces. Records of nine malaria indicators including report submission date, total malaria cases, Plasmodium falciparum case record, Plasmodium vivax case record, clinical malaria, malaria diagnosed with microscopy, malaria diagnosed with (rapid diagnostic test) (RDT), record of drug stocks and records of RDT stocks from 1st January to 31st December 2016 were extracted from the DHIS2 database. The indicators permitted assessment in four core areas: availability, completeness, timeliness and reliability. To explore perceptions and point of view of the stakeholders on the performance of the malaria case reporting system, focus group discussions were conducted with health centre nurses, whilst in-depth interviews were conducted with stakeholder representatives from government (province and national) staff and World Health Organization officials who were users of DHIS2.

Results

Data were extracted from nine health centres in Honoria City Council and 64 health centres in Malaita Province. The completeness and timeliness from the two provinces of all nine indicators were 28.2% and 5.1%, respectively. The most reliable indicator in DHIS2 was ‘clinical malaria’ (i.e. numbers of clinically diagnosed malaria cases) with 62.4% reliability. Challenges to completeness were a lack of supervision, limited feedback, high workload, and a lack of training and refresher courses. Health centres located in geographically remote areas, a lack of regular transport, high workload and too many variables in the reporting forms led to delays in timely reporting. Reliability of reports was impacted by a lack of technical professionals such as statisticians and unavailability of tally sheets and reporting forms.

Conclusion

The availability, completeness, timeliness and reliability of nine malaria indicators collected in DHIS2 were variable within the study area, but generally low. Continued onsite support, supervision, feedback and additional enhancements, such as electronic reporting will be required to further improve the malaria reporting system.

Development of a Predictive Algorithm to Identify Adults With Mobility Limitations Using VA Health Care Administrative Data

An estimated 31.5 million Americans have a mobility limitation. Health care administrative data could be a valuable resource for research on this population but methods for cohort identification are lacking. We developed and tested an algorithm to reliably identify adults with mobility limitation in U.S. Department of Veterans Affairs health care data. We linked diagnosis, encounter, durable medical equipment, and demographic data for 964 veterans to their self-reported mobility limitation from the Medicare Current Beneficiary Survey. We evaluated performance of logistic regression models in classifying mobility limitation. The binary approach (yes/no limitation) had good sensitivity (70%) and specificity (79%), whereas the multilevel approach did not perform well. The algorithms for predicting a binary mobility limitation outcome performed well at discriminating between veterans who did and did not have mobility limitation. Future work should focus on multilevel approaches to predicting mobility limitation and samples with greater proportions of women and younger adults.

Identifying return visits to the Emergency Department: A multi-centre study

BACKGROUND

Patients who return to the Emergency Department (ED) for the same complaint are known to be at risk of adverse events. Monitoring of return visits is considered a way to measure the quality of care provided in the ED, although the most commonly used benchmark of 48h lacks evidence. This study aimed to describe the incidence, characteristics and outcomes of patients with unplanned return visits. The study also aimed to determine the capture rate of the 48-h benchmark using an all-inclusive method of return visit identification.

METHODS

A retrospective cross-sectional study was conducted across three EDs in Sydney, New South Wales from July 1st, 2017 to June 30th, 2018. Visits that occurred within 28 days with the same or similar presenting complaint following discharge from the ED were classified as a return visit. Data were grouped by index and return visit. Descriptive statistics were used to summarise incidence, patient characteristics and outcomes for all presentations. Categorical data were analysed using Chi square tests. Continuous data were analysed using Mann-Whitney when data were not normally distributed and t-tests when normally distributed.

RESULTS

Of all ED presentations (n=164,598), 5860 (3.6%) were identified as a return visit. Return patients were younger than non-return patients, but those that required admission were older (43 vs 33 years, p=<0.01). Abdominal problems were the most common reason for return followed by urological and mental health. The median time to return was 64:51h (IQR 20:35-226:37). Only 43% of return visits occurred within 48h. Return visits to a different ED accounted for 13.2% of return visits.

CONCLUSION

More than half of ED return visits are missed when the existing benchmark of 48h is used. Current policy makers should consider increasing the 48-h benchmark to more accurately reflect the incidence of return visits. Further investigation into the causal factors for return visits is warranted, particularly in patients with abdominal, urological or mental health complaints.

County-Level Socioeconomic Disparities in Use of Medical Services for Management of Infections by Medicare Beneficiaries With Diabetes-United States, 2012

OBJECTIVE

To assess county-level socioeconomic disparities in medical service usage for infections among Medicare beneficiaries with diabetes (MBWDs) who had fee-for-service health insurance claims during 2012.

DESIGN

We used Medicare claims data to calculate percentage of MBWDs with infections.

SETTING

Medicare beneficiaries.

PARTICIPANTS

We estimated the percentage of MBWDs who used medical services for each of 3 groups of infections by sex and quintiles of the prevalence of social factors in the person's county of residence: anatomic site-specific infections; pathogen-specific infections; and HHST infections (human immunodeficiency virus/acquired immunodeficiency syndrome, viral hepatitis, sexually transmitted diseases, and tuberculosis).

MAIN OUTCOME MEASURES

Using quintiles of county-specific socioeconomic determinants, we calculated absolute and relative disparities in each group of infections for male and female MBWDs. We also used regression-based summary measures to estimate the overall average absolute and relative disparities for each infection group.

RESULTS

Of the 4.5 million male MBWDs, 15.8%, 25.3%, and 2.7% had 1 or more site-specific, pathogen-specific, and HHST infections, respectively. Results were similar for females (n = 5.2 million). The percentage of MBWDs with 1 or more infections in each group increased as social disadvantage in the MBWDs' county of residence increased. Absolute and relative county-level socioeconomic disparities in receipt of medical services for 1 or more infections (site- or pathogen-specific) were 12.9 or less percentage points and 65.5% or less, respectively. For HHST infections, percentage of MBWDs having 1 or more HHST infections for persons residing in the highest quintile (Q5) was 3- to 4-fold higher (P < .001) than persons residing in the lowest quintile (Q1).

CONCLUSIONS

Infection burden among MBWDs is generally associated with county-level contextual socioeconomic disadvantage, and the extent of health disparities varies by infection category, socioeconomic factor, and quintiles of socioeconomic disadvantage. The findings imply ongoing need for efforts to identify effective interventions for reducing county-level social disparities in infections among patients with diabetes.

Clinical characteristics and outcomes for paediatric patients admitted with congenital or acquired syphilis: a population-based cohort study

OBJECTIVES

Paediatric congenital and acquired syphilis cases have been increasing since 2012 in the USA. Potential differences in associated hospitalisation trends and healthcare utilisation between the two syphilis entities have not yet been assessed. We sought to compare these entities and describe their clinical characteristics, distribution and impact in the USA.

METHODS

We conducted a population-based cohort study using the 2016 Kids' Inpatient Database (KID) to identify and characterise syphilis-associated hospitalisations among paediatric patients (age 0-21 years) in the USA during the year of 2016. Length of stay and hospitalisation costs for patients with congenital and acquired syphilis were compared in multivariable models.

RESULTS

A total of 1226 hospitalisations with the diagnosis of syphilis were identified. Of these patients, 958 had congenital syphilis and 268 were acquired cases. The mean cost of care for congenital syphilis was $23 644 (SD=1727), while the treatment of a patient with acquired syphilis on average cost $10 749 (SD=1966). Mean length of stay was 8 days greater and mean total costs were $12 895 (US dollars) higher in the congenital syphilis cohort compared with the acquired syphilis cohort. In congenital syphilis, there were greater frequency of cases in the Southern and Western regions of the USA (p<0.001).

CONCLUSION

Congenital syphilis was associated with greater healthcare-related expenditure than acquired syphilis in paediatric patients. In addition to improving patient outcomes, congenital syphilis prevention efforts may significantly reduce healthcare utilisation burden and cost.

Rate of diagnostic errors and serious misdiagnosis-related harms for major vascular events, infections, and cancers: toward a national incidence estimate using the “Big Three”

Background

Missed vascular events, infections, and cancers account for ~75% of serious harms from diagnostic errors. Just 15 diseases from these “Big Three” categories account for nearly half of all serious misdiagnosis-related harms in malpractice claims. As part of a larger project estimating total US burden of serious misdiagnosis-related harms, we performed a focused literature review to measure diagnostic error and harm rates for these 15 conditions.

Methods

We searched PubMed, Google, and cited references. For errors, we selected high-quality, modern, US-based studies, if available, and best available evidence otherwise. For harms, we used literature-based estimates of the generic (disease-agnostic) rate of serious harms (morbidity/mortality) per diagnostic error and applied claims-based severity weights to construct disease-specific rates. Results were validated via expert review and comparison to prior literature that used different methods. We used Monte Carlo analysis to construct probabilistic plausible ranges (PPRs) around estimates.

Results

Rates for the 15 diseases were drawn from 28 published studies representing 91,755 patients. Diagnostic error (false negative) rates ranged from 2.2% (myocardial infarction) to 62.1% (spinal abscess), with a median of 13.6% [interquartile range (IQR) 9.2–24.7] and an aggregate mean of 9.7% (PPR 8.2–12.3). Serious misdiagnosis-related harm rates per incident disease case ranged from 1.2% (myocardial infarction) to 35.6% (spinal abscess), with a median of 5.5% (IQR 4.6–13.6) and an aggregate mean of 5.2% (PPR 4.5–6.7). Rates were considered face valid by domain experts and consistent with prior literature reports.

Conclusions

Diagnostic improvement initiatives should focus on dangerous conditions with higher diagnostic error and misdiagnosis-related harm rates.

Validation of an Optimized Algorithm for Identifying Persons Living With Diagnosed HIV From New York State Medicaid Data, 2006-2014

Algorithms are regularly used to identify persons living with diagnosed human immunodeficiency virus (HIV) (PLWDH) in Medicaid data. To our knowledge, there are no published reports of an HIV algorithm from Medicaid claims codes that have been compared with an HIV surveillance system to assess its sensitivity, specificity, positive predictive value, and negative predictive value in identifying PLWDH. Therefore, our aims in this study were to 1) develop an algorithm that could identify PLWDH in New York State Medicaid data from 2006-2014 and 2) validate this algorithm using the New York State HIV surveillance system. Classification and regression tree analysis identified 16 nodes that we combined to create a case-finding algorithm with 5 criteria. This algorithm identified 86,930 presumed PLWDH, 88.0% of which were verified by matching to the surveillance system. The algorithm yielded a sensitivity of 94.5%, a specificity of 94.4%, a positive predictive value of 88.0%, and a negative predictive value of 97.6%. This validated algorithm has the potential to improve the utility of Medicaid data for assessing health outcomes and programmatic interventions.

Does the Halo Effect for Level 1 Trauma Centers Apply to High-Acuity Nonsurgical Admissions?

CONTEXT

The halo effect describes the improved surgical outcomes at trauma centers for nontrauma conditions.

OBJECTIVE

To determine whether level 1 trauma centers have improved inpatient mortality for common but high-acuity nonsurgical diagnoses (eg, acute myocardial infarction [AMI], congestive heart failure [CHF], and pneumonia [PNA]) compared with non--level 1 trauma centers.

METHODS

The authors conducted a population-based, retrospective cohort study analyzing data from the Healthcare Cost and Utilization Project State Inpatient Database and the American Hospital Association Annual Survey Database. Patients who were admitted with AMI, CHF, and PNA between 2006-2011 in Florida and California were included. Level 1 trauma centers were matched to non-level 1 trauma centers using propensity scoring. The primary outcome was risk-adjusted inpatient mortality for each diagnosis (AMI, CHF, or PNA).

RESULTS

Of the 190,474 patients who were hospitalized for AMI, CHF, or PNA, 94,037 patients (49%) underwent treatment at level 1 trauma centers. The inpatient mortality rates at level 1 trauma centers vs non-level 1 trauma centers for patients with AMI was 8.10% vs 8.40%, respectively (P=.73); for patients with CHF, 2.26% vs 2.71% (P=.90); and for patients with PNA, 2.30% vs 2.70% (P=.25).

CONCLUSION

Level 1 trauma center designation was not associated with improved mortality for high-acuity, nonsurgical medical conditions in this study.

Using standardised patients to assess the quality of medical records: an application and evidence from rural China

BACKGROUND

Medical records play a fundamental role in healthcare delivery, quality assessment and improvement. However, there is little objective evidence on the quality of medical records in low and middle-income countries.

OBJECTIVE

To provide an unbiased assessment of the quality of medical records for outpatient visits to rural facilities in China.

METHODS

A sample of 207 township health facilities across three provinces of China were enrolled. Unannounced standardised patients (SPs) presented to providers following standardised scripts. Three weeks later, investigators returned to collect medical records from each facility. Audio recordings of clinical interactions were then used to evaluate completeness and accuracy of available medical records.

RESULTS

Medical records were located for 210 out of 620 SP visits (33.8%). Of those located, more than 80% contained basic patient information and drug treatment when mentioned in visits, but only 57.6% recorded diagnoses. The most incompletely recorded category of information was patient symptoms (74.3% unrecorded), followed by non-drug treatments (65.2% unrecorded). Most of the recorded information was accurate, but accuracy fell below 80% for some items. The keeping of any medical records was positively correlated with the provider's income (β 0.05, 95% CI 0.01 to 0.09). Providers at hospitals with prescription review were less likely to record completely (β -0.87, 95% CI -1.68 to 0.06). Significant variation by disease type was also found in keeping of any medical record and completeness.

CONCLUSION

Despite the importance of medical records for health system functioning, many rural facilities have yet to implement systems for maintaining patient records, and records are often incomplete when they exist. Prescription review tied to performance evaluation should be implemented with caution as it may create disincentives for record keeping. Interventions to improve record keeping and management are needed.

Mental health of parents of children with a developmental disability in British Columbia, Canada

Background

There is evidence in the literature that parents of children who have a developmental disability experience an increased risk of mental health problems.

Methods

This study used population-level administrative data from the Ministry of Health, British Columbia, Canada, to assess the mental health of parents of children who have a developmental disability compared with the mental health of parents of children who do not have a developmental disability. Population-level and individual explanatory variables available in the data were included in the models.

Results

At a population level, the study found strong evidence that parents of children who have a developmental disability experience higher odds of depression or other mental health diagnoses compared with parents of children who do not have a developmental disability. Age of the parent at birth of the child, income and location of healthcare services were all associated with outcomes.

Conclusion

Parents of children who have a developmental disability may be in need of programmes and services that support their mental health.

Characteristics Associated With Receiving Cataract Surgery in the US Medicare and Veterans Health Administration Populations

Importance

Considerable variation exists with respect to the profiles of patients who receive cataract surgery in the United States.

Objective

To identify patient characteristics associated with receiving cataract surgery within the US Medicare and Veterans Health Administration (VHA) populations.

Design, Setting, and Participants

In this population-based retrospective cohort study of 3 073 465 patients, Medicare and VHA patients with a cataract diagnosis between January 1, 2002, and January 1, 2012, were identified from the 2002-2012 Medicare Part B files (5% sample) and the VHA National Patient Care Database. Patient age, sex, race/ethnicity, region of residence, Charlson Comorbidity Index (CCI) scores, and comorbidities were recorded. Cataract surgery at 1 and 5 years after diagnosis was identified. Data analysis was performed from July 1, 2016, to July 1, 2017.

Main Outcomes and Measures

Odds ratios (ORs) of cataract surgery for selected patient characteristics.

Results

The study sample included 1 156 211 Medicare patients (mean [SD] age, 73.7 [7.0] years) and 1 917 254 VHA patients (mean [SD] age, 66.8 [10.2] years) with a cataract diagnosis. Of the 1 156 211 Medicare patients, 407 103 (35.2%) were 65 to 69 years old, 683 036 (59.1%) were female, and 1 012 670 (87.6%) were white. Of the 1 917 254 VHA patients, 905 455 (47.2%) were younger than 65 years, 1 852 158 (96.6%) were male, and 539 569 (28.1%) were white. A greater proportion of Medicare patients underwent cataract surgery at 1 year (Medicare: 213 589 [18.5%]; VHA: 120 196 [6.3%]) and 5 years (Medicare: 414 586 [35.9%]; VHA: 240 884 [12.6%]) after diagnosis. Factors associated with the greatest odds of surgery at 5 years were older age per 5-year increase (Medicare: OR, 1.24 [95% CI, 1.23-1.24]; VHA: OR, 1.18 [95% CI, 1.17-1.18]), residence in the southern United States vs eastern United States (Medicare: OR, 1.38 [95% CI, 1.36-1.40]; VHA: OR, 1.40 [95% CI, 1.38-1.41]), and presence of chronic pulmonary disease (Medicare: OR, 1.26 [95% CI, 1.24-1.27]; VHA: OR, 1.40 [95% CI, 1.38-1.41]). Within Medicare, female sex was associated with greater odds of surgery at 5 years (OR, 1.14; 95% CI, 1.13-1.15). Higher CCI scores (CCI score ≥3 vs 0-2) were associated with increased odds of surgery among VHA but not Medicare patients at 5 years (Medicare: OR, 0.94 [95% CI, 0.92-0.95]; VHA: OR, 1.24 [95% CI, 1.23-1.36]). Black race vs white race was associated with decreased odds of cataract surgery 5 years after diagnosis (Medicare: OR, 0.79 [95% CI, 0.78-0.81]; VHA: OR, 0.75 [95% CI, 0.73-0.76]).

Conclusions and Relevance

Within both groups, older age, residence in the southern United States, and presence of chronic pulmonary disease were associated with increased odds of cataract surgery. Findings from this study suggest that few disparities exist between the types of patients receiving cataract surgery who are in Medicare vs the VHA, although it is possible that a smaller proportion of VHA patients receive surgery compared with Medicare patients.

Randomised controlled trial of a theory-based intervention to prompt front-line staff to take up the seasonal influenza vaccine

OBJECTIVE

To evaluate the effectiveness of reminder letters informed by social normative theory (a type of 'nudge theory') on uptake of seasonal influenza vaccination by front-line hospital staff.

DESIGN

Individually randomised controlled trial.

SETTING

A large acute care hospital in England.

PARTICIPANTS

Front-line staff employed by the hospital (n=7540) were randomly allocated to one of four reminder types in a factorial design.

INTERVENTIONS

The standard letter included only general information directing the staff to take up the vaccine. A second letter highlighted a type of social norm based on peer comparisons. A third letter highlighted a type of social norm based on an appeal to authority. A fourth letter included a combination of the social norms.

MAIN OUTCOME MEASURE

The proportion of hospital staff vaccinated on-site.

RESULTS

Vaccine coverage was 43% (812/1885) in the standard letter group, 43% (818/1885) in the descriptive norms group, 43% (814/1885) in the injunctive norms group and 43% (812/1885) in the combination group. There were no statistically significant effects of either norm or the interaction. The OR for the descriptive norms factor is 1.01 (0.89-1.15) in the absence of the injunctive norms factor and 1.00 (0.88-1.13) in its presence. The OR for the injunctive norms factor is 1.00 (0.88-1.14) in the absence of the descriptive norms factor and 0.99 (0.87-1.12) in its presence.

CONCLUSIONS

We find no evidence that the uptake of the seasonal influenza vaccination is affected by reminders using social norms to motivate uptake.

Randomised controlled trial of a theory-based intervention to prompt front-line staff to take up the seasonal influenza vaccine

Objective

To evaluate the effectiveness of reminder letters informed by social normative theory (a type of ‘nudge theory’) on uptake of seasonal influenza vaccination by front-line hospital staff.

Design

Individually randomised controlled trial.

Setting

A large acute care hospital in England.

Participants

Front-line staff employed by the hospital (n=7540) were randomly allocated to one of four reminder types in a factorial design.

Interventions

The standard letter included only general information directing the staff to take up the vaccine. A second letter highlighted a type of social norm based on peer comparisons. A third letter highlighted a type of social norm based on an appeal to authority. A fourth letter included a combination of the social norms.

Main outcome measure

The proportion of hospital staff vaccinated on-site.

Results

Vaccine coverage was 43% (812/1885) in the standard letter group, 43% (818/1885) in the descriptive norms group, 43% (814/1885) in the injunctive norms group and 43% (812/1885) in the combination group. There were no statistically significant effects of either norm or the interaction. The OR for the descriptive norms factor is 1.01 (0.89–1.15) in the absence of the injunctive norms factor and 1.00 (0.88–1.13) in its presence. The OR for the injunctive norms factor is 1.00 (0.88–1.14) in the absence of the descriptive norms factor and 0.99 (0.87–1.12) in its presence.

Conclusions

We find no evidence that the uptake of the seasonal influenza vaccination is affected by reminders using social norms to motivate uptake.

Administrative database as a source for assessment of systemic lupus erythematosus prevalence: Estonian experience

Background

Administrative database research is widely applied in the field of epidemiology. However, the results of the studies depend on the type of database used and the algorithms applied for case ascertainment. The optimal methodology for identifying patients with rheumatic diseases from administrative databases is yet not known. Our aim was to describe an administrative database as a source for estimation of epidemiological characteristics on an example of systemic lupus erythematosus (SLE, ICD-10 code M32) prevalence assessment in the database of the Estonian Health Insurance Fund (EHIF).

Methods

Code M32 billing episodes were extracted from the EHIF database 2006–2010. For all cases where M32 was assigned by a rheumatologist less than four times during the study period, diagnosis verification process using health care providers’ (HCP) databases was applied. For M32 cases assigned by a rheumatologist four times or more, diagnoses were verified for a randomly selected sample.

Results

From 677 persons with code M32 assigned in EHIF database, 404 were demonstrated having “true SLE”. The code M32 positive predictive value (PPV) for the whole EHIF database was 60%; PPV varies remarkably by specialty of a physician and repetition of the code assignment. The false positive M32 codes were predominantly initial diagnoses which were not confirmed afterwards; in many cases, a rheumatic condition other than SLE was later diagnosed.

Conclusions

False positive codes due to tentative diagnoses may be characteristic for conditions with a complicated diagnosis process like SLE and need to be taken into account when performing administrative database research.

PheValuator: Development and evaluation of a phenotype algorithm evaluator

BACKGROUND

The primary approach for defining disease in observational healthcare databases is to construct phenotype algorithms (PAs), rule-based heuristics predicated on the presence, absence, and temporal logic of clinical observations. However, a complete evaluation of PAs, i.e., determining sensitivity, specificity, and positive predictive value (PPV), is rarely performed. In this study, we propose a tool (PheValuator) to efficiently estimate a complete PA evaluation.

METHODS

We used 4 administrative claims datasets: OptumInsight's de-identified Clinformatics™ Datamart (Eden Prairie,MN); IBM MarketScan Multi-State Medicaid); IBM MarketScan Medicare Supplemental Beneficiaries; and IBM MarketScan Commercial Claims and Encounters from 2000 to 2017. Using PheValuator involves (1) creating a diagnostic predictive model for the phenotype, (2) applying the model to a large set of randomly selected subjects, and (3) comparing each subject's predicted probability for the phenotype to inclusion/exclusion in PAs. We used the predictions as a 'probabilistic gold standard' measure to classify positive/negative cases. We examined 4 phenotypes: myocardial infarction, cerebral infarction, chronic kidney disease, and atrial fibrillation. We examined several PAs for each phenotype including 1-time (1X) occurrence of the diagnosis code in the subject's record and 1-time occurrence of the diagnosis in an inpatient setting with the diagnosis code as the primary reason for admission (1X-IP-1stPos).

RESULTS

Across phenotypes, the 1X PA showed the highest sensitivity/lowest PPV among all PAs. 1X-IP-1stPos yielded the highest PPV/lowest sensitivity. Specificity was very high across algorithms. We found similar results between algorithms across datasets.

CONCLUSION

PheValuator appears to show promise as a tool to estimate PA performance characteristics.

A simulation-based approach to training in heuristic clinical decision-making

Background

Cognitive biases may negatively impact clinical decision-making. The dynamic nature of a simulation environment can facilitate heuristic decision-making which can serve as a teaching opportunity.

Methods

Momentum bias, confirmation bias, playing-the-odds bias, and order-effect bias were integrated into four simulation scenarios. Clinical simulation educators and human factors specialists designed a script of events during scenarios to trigger heuristic decision-making. Debriefing included the exploration of frames (mental models) resulting in the observed actions, as well as a discussion of specific bias-prone frames and bias-resistant frames. Simulation sessions and debriefings were coded to measure the occurrence of bias, recovery from biased decision-making, and effectiveness of debriefings.

Results

Twenty medical residents and 18 medical students participated in the study. Twenty pairs (of one medical student and one resident) and two individuals (medical residents alone) completed a simulation session. Evidence of bias was observed in 11 of 20 (55%) sessions. While most participant pairs were able to avoid or recover from the anticipated bias, there were three sessions with no recovery. Evaluation of debriefings showed exploration of frames in all the participant pairs. Establishing new bias-resistant frames occurred more often when the learners experienced the bias.

Conclusions

Instructional design using experiential learning can focus learner attention on the specific elements of diagnostic decision-making. Using scenario design and debriefing enabled trainees to experience and analyze their own cognitive biases.

Development and validation of an electronic medical record (EMR)-based computed phenotype of HIV-1 infection

Background

Electronic medical record (EMR) computed algorithms allow investigators to screen thousands of patient records to identify specific disease cases. No computed algorithms have been developed to detect all cases of human immunodeficiency virus (HIV) infection using administrative, laboratory, and clinical documentation data outside of the Veterans Health Administration. We developed novel EMR-based algorithms for HIV detection and validated them in a cohort of subjects in the Duke University Health System (DUHS).

Methods

We created 2 novel algorithms to identify HIV-infected subjects. Algorithm 1 used laboratory studies and medications to identify HIV-infected subjects, whereas Algorithm 2 used International Classification of Diseases, Ninth Revision (ICD-9) codes, medications, and laboratory testing. We applied the algorithms to a well-characterized cohort of patients and validated both against the gold standard of physician chart review. We determined sensitivity, specificity, and prevalence of HIV between 2007 and 2011 in patients seen at DUHS.

Results

A total of 172 271 patients were detected with complete data; 1063 patients met algorithm criteria for HIV infection. In all, 970 individuals were identified by both algorithms, 78 by Algorithm 1 alone, and 15 by Algorithm 2 alone. The sensitivity and specificity of each algorithm were 78% and 99%, respectively, for Algorithm 1 and 77% and 100% for Algorithm 2. The estimated prevalence of HIV infection at DUHS between 2007 and 2011 was 0.6%.

Conclusions

EMR-based phenotypes of HIV infection are capable of detecting cases of HIV-infected adults with good sensitivity and specificity. These algorithms have the potential to be adapted to other EMR systems, allowing for the creation of cohorts of patients across EMR systems.

Comparative Accuracy of Diagnosis by Collective Intelligence of Multiple Physicians vs Individual Physicians

IMPORTANCE

The traditional approach of diagnosis by individual physicians has a high rate of misdiagnosis. Pooling multiple physicians' diagnoses (collective intelligence) is a promising approach to reducing misdiagnoses, but its accuracy in clinical cases is unknown to date.

OBJECTIVE

To assess how the diagnostic accuracy of groups of physicians and trainees compares with the diagnostic accuracy of individual physicians.

DESIGN, SETTING, AND PARTICIPANTS

Cross-sectional study using data from the Human Diagnosis Project (Human Dx), a multicountry data set of ranked differential diagnoses by individual physicians, graduate trainees, and medical students (users) solving user-submitted, structured clinical cases. From May 7, 2014, to October 5, 2016, groups of 2 to 9 randomly selected physicians solved individual cases. Data analysis was performed from March 16, 2017, to July 30, 2018.

MAIN OUTCOMES AND MEASURES

The primary outcome was diagnostic accuracy, assessed as a correct diagnosis in the top 3 ranked diagnoses for an individual; for groups, the top 3 diagnoses were a collective differential generated using a weighted combination of user diagnoses with a variety of approaches. A version of the McNemar test was used to account for clustering across repeated solvers to compare diagnostic accuracy.

RESULTS

Of the 2069 users solving 1572 cases from the Human Dx data set, 1228 (59.4%) were residents or fellows, 431 (20.8%) were attending physicians, and 410 (19.8%) were medical students. Collective intelligence was associated with increasing diagnostic accuracy, from 62.5% (95% CI, 60.1%-64.9%) for individual physicians up to 85.6% (95% CI, 83.9%-87.4%) for groups of 9 (23.0% difference; 95% CI, 14.9%-31.2%; P < .001). The range of improvement varied by the specifications used for combining groups' diagnoses, but groups consistently outperformed individuals regardless of approach. Absolute improvement in accuracy from individuals to groups of 9 varied by presenting symptom from an increase of 17.3% (95% CI, 6.4%-28.2%; P = .002) for abdominal pain to 29.8% (95% CI, 3.7%-55.8%; P = .02) for fever. Groups from 2 users (77.7% accuracy; 95% CI, 70.1%-84.6%) to 9 users (85.5% accuracy; 95% CI, 75.1%-95.9%) outperformed individual specialists in their subspecialty (66.3% accuracy; 95% CI, 59.1%-73.5%; P < .001 vs groups of 2 and 9).

CONCLUSIONS AND RELEVANCE

A collective intelligence approach was associated with higher diagnostic accuracy compared with individuals, including individual specialists whose expertise matched the case diagnosis, across a range of medical cases. Given the few proven strategies to address misdiagnosis, this technique merits further study in clinical settings.

Assessing the quality of primary healthcare in seven Chinese provinces with unannounced standardised patients: protocol of a cross-sectional survey

INTRODUCTION

Primary healthcare (PHC) serves as the cornerstone for the attainment of universal health coverage (UHC). Efforts to promote UHC should focus on the expansion of access and on healthcare quality. However, robust quality evidence has remained scarce in China. Common quality assessment methods such as chart abstraction, patient rating and clinical vignette use indirect information that may not represent real practice. This study will send standardised patients (SP or healthy person trained to consistently simulate the medical history, physical symptoms and emotional characteristics of a real patient) unannounced to PHC providers to collect quality information and represent real practice.

METHODS AND ANALYSIS

1981 SP-clinician visits will be made to a random sample of PHC providers across seven provinces in China. SP cases will be developed for 10 tracer conditions in PHC. Each case will include a standard script for the SP to use and a quality checklist that the SP will complete after the clinical visit to indicate diagnostic and treatment activities performed by the clinician. Patient-centredness will be assessed according to the Patient Perception of Patient-Centeredness Rating Scale by the SP. SP cases and the checklist will be developed through a standard protocol and assessed for content, face and criterion validity, and test-retest and inter-rater reliability before its full use. Various descriptive analyses will be performed for the survey results, such as a tabulation of quality scores across geographies and provider types.

ETHICS AND DISSEMINATION

This study has been reviewed and approved by the Institutional Review Board of the School of Public Health of Sun Yat-sen University (#SYSU 2017-011). Results will be actively disseminated through print and social media, and SP tools will be made available for other researchers.

Quality of patient information in emergency department

PURPOSE

Data comprise one of the key resources currently used in organizations. High-quality data are those that are appropriate for use by the customer. The quality of data is a key factor in determining the level of healthcare in hospitals, and its improvement leads to an improved quality of health and treatment and ultimately increases patient satisfaction. The purpose of this paper is to assess the quality of emergency patients' information in a hospital information system.

DESIGN/METHODOLOGY/APPROACH

This cross-sectional study was conducted on 385 randomly selected records of patients admitted to the emergency department of Shahid Beheshti Hospital in Kashan, Iran, in 2016. Data on five dimensions of quality, including accuracy, accessibility, timeliness, completeness and definition, were collected using a researcher-made checklist and were then analyzed in SPSS. The results are presented using descriptive statistics, such as frequency distribution and percentage.

FINDINGS

The overall quality of emergency patients' information in the hospital information system was 86 percent, and the dimensions of quality scored 87.7 percent for accuracy, 86.8 percent for completeness, 83.9 percent for timeliness, 79 percent for definition and 62.1 percent for accessibility.

ORIGINALITY/VALUE

Increasing the quality of patient information at emergency departments can lead to improvements in the timely diagnosis and management of diseases and patient and personnel satisfaction, and reduce hospital costs.

Do Diagnostic and Procedure Codes Within Population-Based, Administrative Datasets Accurately Identify Patients with Rectal Cancer?

BACKGROUND

Procedural and diagnostic codes may inaccurately identify specific patient populations within administrative datasets.

PURPOSE

Measure the accuracy of previously used coding algorithms using administrative data to identify patients with rectal cancer resections (RCR).

METHODS

Using a previously published coding algorithm, we re-created a RCR cohort within administrative databases, limiting the search to a single institution. The accuracy of this cohort was determined against a gold standard reference population. A systematic review of the literature was then performed to identify studies that use similar coding methods to identify RCR cohorts and whether or not they comment on accuracy.

RESULTS

Over the course of the study period, there were 664,075 hospitalizations at our institution. Previously used coding algorithms identified 1131 RCRs (administrative data incidence 1.70 per 1000 hospitalizations). The gold standard reference population was 821 RCR over the same period (1.24 per 1000 hospitalizations). Administrative data methods yielded a RCR cohort of moderate accuracy (sensitivity 89.5%, specificity 99.9%) and poor positive predictive value (64.9%). Literature search identified 18 studies that utilized similar coding methods to derive a RCR cohort. Only 1/18 (5.6%) reported on the accuracy of their study cohort.

CONCLUSIONS

The use of diagnostic and procedure codes to identify RCR within administrative datasets may be subject to misclassification bias because of low PPV. This underscores the importance of reporting on the accuracy of RCR cohorts derived within population-based datasets.

Does place of service matter? A utilisation and cost analysis of sexually transmissible infection testing from 2012 claims data

Background In this study, a previous study on the utilisation and cost of sexually transmissible infection (STI) tests was augmented by focusing on outpatient place of service for the most utilised tests.

METHODS

Claims for eight STI tests [chlamydia, gonorrhoea, hepatitis B virus (HBV), HIV, human papillomavirus (HPV), herpes simplex virus type 2 (HSV2), syphilis and trichomoniasis] using the most utilised current procedural terminology (CPT) code for each STI from the 2012 MarketScan outpatient table were extracted. The volume and costs by gender and place of service were then summarised. Finally, semi-log regression analyses were used to further examine and compare costs.

RESULTS

Females had a higher number of test claims than males in all places of service for each STI. Together, claims from 'Independent Laboratories', 'Office' and 'Outpatient hospital' accounted for over 93% of all the test claims. The cost of tests were slightly (<5%) different between males and females for most places of service. Except for the estimated average cost for 'Outpatient hospital', the estimated average costs for the other categories were significantly lower (15-80%, P<0.01) than the estimated average cost for 'Emergency Room - Hospital' for all the STIs. Among the predominant service venues, test costs from 'Independent Laboratory' and 'Office' were 30% to 69% lower (P<0.01) than those from 'Outpatient Hospital'.

CONCLUSIONS

Even though the results from this study are not generalisable, our study shows that almost all STI tests from outpatient claims data were performed in three service venues with considerable cost variations.

Can Text-Search Methods of Pathology Reports Accurately Identify Patients with Rectal Cancer in Large Administrative Databases?

Background:

The aim of this study is to derive and to validate a cohort of rectal cancer surgical patients within administrative datasets using text-search analysis of pathology reports.

Materials and Methods:

A text-search algorithm was developed and validated on pathology reports from 694 known rectal cancers, 1000 known colon cancers, and 1000 noncolorectal specimens. The algorithm was applied to all pathology reports available within the Ottawa Hospital Data Warehouse from 1996 to 2010. Identified pathology reports were validated as rectal cancer specimens through manual chart review. Sensitivity, specificity, and positive predictive value (PPV) of the text-search methodology were calculated.

Results:

In the derivation cohort of pathology reports (n = 2694), the text-search algorithm had a sensitivity and specificity of 100% and 98.6%, respectively. When this algorithm was applied to all pathology reports from 1996 to 2010 (n = 284,032), 5588 pathology reports were identified as consistent with rectal cancer. Medical record review determined that 4550 patients did not have rectal cancer, leaving a final cohort of 1038 rectal cancer patients. Sensitivity and specificity of the text-search algorithm were 100% and 98.4%, respectively. PPV of the algorithm was 18.6%.

Conclusions:

Text-search methodology is a feasible way to identify all rectal cancer surgery patients through administrative datasets with high sensitivity and specificity. However, in the presence of a low pretest probability, text-search methods must be combined with a validation method, such as manual chart review, to be a viable approach.

Incident learning in radiation oncology: A review

Incident learning is a key component for maintaining safety and quality in healthcare. Its use is well established and supported by professional society recommendations, regulations and accreditation, and objective evidence. There is an active interest in incident learning systems (ILS) in radiation oncology, with over 40 publications since 2010. This article is intended as a comprehensive topic review of ILS in radiation oncology, including history and summary of existing literature, nomenclature and categorization schemas, operational aspects of ILS at the institutional level including event handling and root cause analysis, and national and international ILS for shared learning. Core principles of patient safety in the context of ILS are discussed, including the systems view of error, culture of safety, and contributing factors such as cognitive bias. Finally, the topics of medical error disclosure and second victim syndrome are discussed. In spite of the rapid progress and understanding of ILS, challenges remain in applying ILS to the radiation oncology context. This comprehensive review may serve as a springboard for further work.

India's Largest Hospital Insurance Program Faces Challenges In Using Claims Data To Measure Quality

The routine data generated by India’s universal coverage programs offer an important opportunity to evaluate and track the quality of health care systematically and on a large scale. We examined the potential and challenges of measuring the quality of hospital care through claims data from India’s hospital insurance program for the poor, Rashtriya Swasthya Bima Yojana (RSBY). Using data from one district in India, we illustrate how these data already provide useful insights and show that simple efforts to enhance data quality and an effort to expand the data captured could facilitate RSBY’s ability to track quality of care. The data collected by RSBY has significant potential to characterize and uncover the provision of low-quality care and help inform much-needed efforts to raise the quality of hospital care.