ResearchPad - Health Informatics https://www.researchpad.co Default RSS Feed en-us © 2020 Newgen KnowledgeWorks <![CDATA[Evaluation of a Deep Neural Network for Automated Classification of Colorectal Polyps on Histopathologic Slides]]> https://www.researchpad.co/product?articleinfo=N2f55895c-985c-4d6c-bab1-9e029c75ba88

This prognostic study evaluates the performance and generalizability of a deep neural network trained on data from a single institution for classification of colorectal polyps on histopathologic slide images.

]]>
<![CDATA[Antibiotic use for Australian Aboriginal children in three remote Northern Territory communities]]> https://www.researchpad.co/product?articleinfo=N999fa4e6-a15c-456a-862e-2e1ce88316a9

Objective

To describe antibiotic prescription rates for Australian Aboriginal children aged <2 years living in three remote Northern Territory communities.

Design

A retrospective cohort study using electronic health records.

Setting

Three primary health care centres located in the Katherine East region.

Participants

Consent was obtained from 149 mothers to extract data from 196 child records. There were 124 children born between January 2010 and July 2014 who resided in one of the three chosen communities and had electronic health records for their first two years of life.

Main outcome measures

Antibiotic prescription rates, factors associated with antibiotic prescription and factors associated with appropriate antibiotic prescription.

Results

There were 5,675 Primary Health Care (PHC) encounters for 124 children (median 41, IQR 25.5, 64). Of the 5,675 PHC encounters, 1,542 (27%) recorded at least one infection (total 1,777) and 1,330 (23%) had at least one antibiotic prescription recorded (total 1,468). Children had a median five (IQR 2, 9) prescriptions in both their first and second year of life, with a prescription rate of 5.99/person year (95% CI 5.35, 6.63). Acute otitis media was the most common infection (683 records, 38%) and Amoxycillin was the most commonly prescribed antibiotic (797 prescriptions, 54%). Of the 1,468 recorded prescriptions, 398 (27%) had no infection recorded and 116 (8%) with an infection recorded were not aligned with local treatment guidelines.

Conclusion

Prescription rates for Australian Aboriginal children in these communities are significantly higher than that reported nationally for non-Aboriginal Australians. Prescriptions predominantly aligned with treatment guidelines in this setting where there is a high burden of infectious disease.

]]>
<![CDATA[Reporting quality of studies using machine learning models for medical diagnosis: a systematic review]]> https://www.researchpad.co/product?articleinfo=N70254039-1fb6-4d74-aa07-94d45a574ede

Aims

We conducted a systematic review assessing the reporting quality of studies validating models based on machine learning (ML) for clinical diagnosis, with a specific focus on the reporting of information concerning the participants on which the diagnostic task was evaluated on.

Method

Medline Core Clinical Journals were searched for studies published between July 2015 and July 2018. Two reviewers independently screened the retrieved articles, a third reviewer resolved any discrepancies. An extraction list was developed from the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis guideline. Two reviewers independently extracted the data from the eligible articles. Third and fourth reviewers checked, verified the extracted data as well as resolved any discrepancies between the reviewers.

Results

The search results yielded 161 papers, of which 28 conformed to the eligibility criteria. Detail of data source was reported in 24 of the 28 papers. For all of the papers, the set of patients on which the ML-based diagnostic system was evaluated was partitioned from a larger dataset, and the method for deriving such set was always reported. Information on the diagnostic/non-diagnostic classification was reported well (23/28). The least reported items were the use of reporting guideline (0/28), distribution of disease severity (8/28 patient flow diagram (10/28) and distribution of alternative diagnosis (10/28). A large proportion of studies (23/28) had a delay between the conduct of the reference standard and ML tests, while one study did not and four studies were unclear. For 15 studies, it was unclear whether the evaluation group corresponded to the setting in which the ML test will be applied to.

Conclusion

All studies in this review failed to use reporting guidelines, and a large proportion of them lacked adequate detail on participants, making it difficult to replicate, assess and interpret study findings.

PROSPERO registration number

CRD42018099167.

]]>
<![CDATA[Association of the Meaningful Use Electronic Health Record Incentive Program With Health Information Technology Venture Capital Funding]]> https://www.researchpad.co/product?articleinfo=N72482149-cc43-4720-bacb-2865c8adc30b

Key Points

Question

What is the association between the Health Information Technology for Economic and Clinical Health Act, a federal subsidy program for electronic health record adoption, and venture capital investments and innovation in health care information technology?

Findings

This economic evaluation used observational data on venture capital activity in the US and a difference-in-differences design and found that investments in health care information technology and electronic health record companies increased at a much faster rate than venture capital investments as a whole and that these investments were more likely to be seed-stage (very early) investments compared with all other industries and 13.6% more likely to be seed stage compared with non–information technology health care investments.

Meaning

The Health Information Technology for Economic and Clinical Health Act’s incentive program was associated with increased investment in health care information technology and electronic health record–related companies compared with other industries, with an emphasis on early-stage companies, suggesting an important role for incentives in promoting innovation.

]]>
<![CDATA[Assessing mental health service user and carer involvement in physical health care planning: The development and validation of a new patient-reported experience measure]]> https://www.researchpad.co/product?articleinfo=5c6dc9a5d5eed0c484529f71

Background

People living with serious mental health conditions experience increased morbidity due to physical health issues driven by medication side-effects and lifestyle factors. Coordinated mental and physical healthcare delivered in accordance with a care plan could help to reduce morbidity and mortality in this population. Efforts to develop new models of care are hampered by a lack of validated instruments to accurately assess the extent to which mental health services users and carers are involved in care planning for physical health.

Objective

To develop a brief and accurate patient-reported experience measure (PREM) capable of assessing involvement in physical health care planning for mental health service users and their carers.

Methods

We employed psychometric and statistical techniques to refine a bank of candidate questionnaire items, derived from qualitative interviews, into a valid and reliable measure involvement in physical health care planning. We assessed the psychometric performance of the item bank using modern psychometric analyses. We assessed unidimensionality, scalability, fit to the partial credit Rasch model, category threshold ordering, local dependency, differential item functioning, and test-retest reliability. Once purified of poorly performing and erroneous items, we simulated computerized adaptive testing (CAT) with 15, 10 and 5 items using the calibrated item bank.

Results

Issues with category threshold ordering, local dependency and differential item functioning were evident for a number of items in the nascent item bank and were resolved by removing problematic items. The final 19 item PREM had excellent fit to the Rasch model fit (x2 = 192.94, df = 1515, P = .02, RMSEA = .03 (95% CI = .01-.04). The 19-item bank had excellent reliability (marginal r = 0.87). The correlation between questionnaire scores at baseline and 2-week follow-up was high (r = .70, P < .01) and 94.9% of assessment pairs were within the Bland Altman limits of agreement. Simulated CAT demonstrated that assessments could be made using as few as 10 items (mean SE = .43).

Discussion

We developed a flexible patient reported outcome measure to quantify service user and carer involvement in physical health care planning. We demonstrate the potential to substantially reduce assessment length whilst maintaining reliability by utilizing CAT.

]]>
<![CDATA[Trazodone use and risk of dementia: A population-based cohort study]]> https://www.researchpad.co/product?articleinfo=5c63396bd5eed0c484ae66ce

Background

In vitro and animal studies have suggested that trazodone, a licensed antidepressant, may protect against dementia. However, no studies have been conducted to assess the effect of trazodone on dementia in humans. This electronic health records study assessed the association between trazodone use and the risk of developing dementia in clinical practice.

Methods and findings

The Health Improvement Network (THIN), an archive of anonymised medical and prescribing records from primary care practices in the United Kingdom, contains records of over 15 million patients. We assessed patients from THIN aged ≥50 years who received at least two consecutive prescriptions for an antidepressant between January 2000 and January 2017. We compared the risk of dementia among patients who were prescribed trazodone to that of patients with similar baseline characteristics prescribed other antidepressants, using a Cox regression model with 1:5 propensity score matching. Patients prescribed trazodone who met the inclusion criteria (n = 4,716; 59.2% female) were older (mean age 70.9 ± 13.1 versus 65.6 ± 11.4 years) and were more likely than those prescribed other antidepressants (n = 420,280; 59.7% female) to have cerebrovascular disease and use anxiolytic or antipsychotic drugs. After propensity score matching, 4,596 users of trazadone and 22,980 users of other antidepressants were analysed. The median time to dementia diagnosis for people prescribed trazodone was 1.8 years (interquartile range [IQR] = 0.5–5.0 years). Incidence of dementia among patients taking trazodone was higher than in matched users of other antidepressants (1.8 versus 1.1 per 100 person-years), with a hazard ratio (HR) of 1.80 (95% confidence interval [CI] 1.56–2.09; p < 0.001). However, our results do not suggest a causal association. When we restricted the control group to users of mirtazapine only in a sensitivity analysis, the findings were very similar to the results of the main analysis. The main limitation of our study is the possibility of indication bias, because people in the prodromal stage of dementia might be preferentially prescribed trazodone. Due to the observational nature of this study, we cannot rule out residual confounding.

Conclusions

In this study of UK population-based electronic health records, we found no association between trazodone use and a reduced risk of dementia compared with other antidepressants. These results suggest that the clinical use of trazodone is not associated with a reduced risk of dementia.

]]>
<![CDATA[More than one in three proxies do not know their loved one’s current code status: An observational study in a Maryland ICU]]> https://www.researchpad.co/product?articleinfo=5c5b5258d5eed0c4842bc6a8

Rationale

The majority of ICU patients lack decision-making capacity at some point during their ICU stay. However the extent to which proxy decision-makers are engaged in decisions about their patient’s care is challenging to quantify.

Objectives

To assess 1)whether proxies know their patient’s actual code status as recorded in the electronic medical record (EMR), and 2)whether code status orders reflect ICU patient preferences as reported by proxy decision-makers.

Methods

We enrolled proxy decision-makers for 96 days starting January 4, 2016. Proxies were asked about the patient’s goals of care, preferred code status, and actual code status. Responses were compared to code status orders in the EMR at the time of interview. Characteristics of patients and proxies who correctly vs incorrectly identified actual code status were compared, as were characteristics of proxies who reported a preferred code status that did vs did not match actual code status.

Measurements and main results

Among 111 proxies, 42 (38%) were incorrect or unsure about the patient’s actual code status and those who were correct vs. incorrect or unsure were similar in age, race, and years of education (P>0.20 for all comparisons). Twenty-nine percent reported a preferred code status that did not match the patient’s code status in the EMR. Matching preferred and actual code status was not associated with a patient’s age, gender, income, admission diagnosis, or subsequent in-hospital mortality or with proxy age, gender, race, education level, or relation to the patient (P>0.20 for all comparisons).

Conclusions

More than 1 in 3 proxies is incorrect or unsure about their patient’s actual code status and more than 1 in 4 proxies reported that a preferred code status that did not match orders in the EMR. Proxy age, race, gender and education level were not associated with correctly identifying code status or code status concordance.

]]>
<![CDATA[District-level health management and health system performance]]> https://www.researchpad.co/product?articleinfo=5c5df34ad5eed0c4845810c0

Strengthening district-level management may be an important lever for improving key public health outcomes in low-income settings; however, previous studies have not established the statistical associations between better management and primary healthcare system performance in such settings. To explore this gap, we conducted a cross-sectional study of 36 rural districts and 226 health centers in Ethiopia, a country which has made ambitious investment in expanding access to primary care over the last decade. We employed quantitative measure of management capacity at both the district health office and health center levels and used multiple regression models, accounting for clustering of health centers within districts, to estimate the statistical association between management capacity and a key performance indicator (KPI) summary score based on antenatal care coverage, contraception use, skilled birth attendance, infant immunization, and availability of essential medications. In districts with above median district management capacity, health center management capacity was strongly associated (p < 0.05) with KPI performance. In districts with below median management capacity, health center management capacity was not associated with KPI performance. Having more staff at the district health office was also associated with better KPI performance (p < 0.05) but only in districts with above median management capacity. The results suggest that district-level management may provide an opportunity for improving health system performance in low-income country settings.

]]>
<![CDATA[Assessment of US Hospital Compliance With Regulations for Patients’ Requests for Medical Records]]> https://www.researchpad.co/product?articleinfo=5c50f63fd5eed0c48462c160

Importance

Although federal law has long promoted patients’ access to their protected health information, this access remains limited. Previous studies have demonstrated some issues in requesting release of medical records, but, to date, there has been no comprehensive review of the challenges that exist in all aspects of the request process.

Objective

To evaluate the current state of medical records request processes of US hospitals in terms of compliance with federal and state regulations and ease of patient access.

Design, Setting, and Participants

A cross-sectional study of medical records request processes was conducted between August 1 and December 7, 2017, in 83 top-ranked US hospitals with independent medical records request processes and medical records departments reachable by telephone. Hospitals were ranked as the top 20 hospitals for each of the 16 adult specialties in the 2016-2017 US News & World Report Best Hospitals National Rankings.

Exposures

Scripted interview with medical records departments in a single-blind, simulated patient experience.

Main Outcomes and Measures

Requestable information (entire medical record, laboratory test results, medical history and results of physical examination, discharge summaries, consultation reports, physician orders, and other), formats of release (pick up in person, mail, fax, email, CD, and online patient portal), costs, and request processing times, identified on medical records release authorization forms and through telephone calls with medical records departments.

Results

Among the 83 top-ranked US hospitals representing 29 states, there was discordance between information provided on authorization forms and that obtained from the simulated patient telephone calls in terms of requestable information, formats of release, and costs. On the forms, as few as 9 hospitals (11%) provided the option of selecting 1 of the categories of information and only 44 hospitals (53%) provided patients the option to acquire the entire medical record. On telephone calls, all 83 hospitals stated that they were able to release entire medical records to patients. There were discrepancies in information given in telephone calls vs on the forms between the formats hospitals stated that they could use to release information (69 [83%] vs 40 [48%] for pick up in person, 20 [24%] vs 14 [17%] for fax, 39 [47%] vs 27 [33%] for email, 55 [66%] vs 35 [42%] for CD, and 21 [25%] vs 33 [40%] for online patient portals), additionally demonstrating noncompliance with federal regulations in refusing to provide records in the format requested by the patient. There were 48 hospitals that had costs of release (as much as $541.50 for a 200-page record) above the federal recommendation of $6.50 for electronically maintained records. At least 6 of the hospitals (7%) were noncompliant with state requirements for processing times.

Conclusions and Relevance

The study revealed that there are discrepancies in the information provided to patients regarding the medical records release processes and noncompliance with federal and state regulations and recommendations. Policies focused on improving patient access may require stricter enforcement to ensure more transparent and less burdensome medical records request processes for patients.

]]>
<![CDATA[Value of Neighborhood Socioeconomic Status in Predicting Risk of Outcomes in Studies That Use Electronic Health Record Data]]> https://www.researchpad.co/product?articleinfo=5c50f5fbd5eed0c48462bb8b

Importance

Data from electronic health records (EHRs) are increasingly used for risk prediction. However, EHRs do not reliably collect sociodemographic and neighborhood information, which has been shown to be associated with health. The added contribution of neighborhood socioeconomic status (nSES) in predicting health events is unknown and may help inform population-level risk reduction strategies.

Objective

To quantify the association of nSES with adverse outcomes and the value of nSES in predicting the risk of adverse outcomes in EHR-based risk models.

Design, Setting, and Participants

Cohort study in which data from 90 097 patients 18 years or older in the Duke University Health System and Lincoln Community Health Center EHR from January 1, 2009, to December 31, 2015, with at least 1 health care encounter and residence in Durham County, North Carolina, in the year prior to the index date were linked with census tract data to quantify the association between nSES and the risk of adverse outcomes. Machine learning methods were used to develop risk models and determine how adding nSES to EHR data affects risk prediction. Neighborhood socioeconomic status was defined using the Agency for Healthcare Research and Quality SES index, a weighted measure of multiple indicators of neighborhood deprivation.

Main Outcomes and Measures

Outcomes included use of health care services (emergency department and inpatient and outpatient encounters) and hospitalizations due to accidents, asthma, influenza, myocardial infarction, and stroke.

Results

Among the 90 097 patients in the training set of the study (57 507 women and 32 590 men; mean [SD] age, 47.2 [17.7] years) and the 122 812 patients in the testing set of the study (75 517 women and 47 295 men; mean [SD] age, 46.2 [17.9] years), those living in neighborhoods with lower nSES had a shorter time to use of emergency department services and inpatient encounters, as well as a shorter time to hospitalizations due to accidents, asthma, influenza, myocardial infarction, and stroke. The predictive value of nSES varied by outcome of interest (C statistic ranged from 0.50 to 0.63). When added to EHR variables, nSES did not improve predictive performance for any health outcome.

Conclusions and Relevance

Social determinants of health, including nSES, are associated with the health of a patient. However, the results of this study suggest that information on nSES may not contribute much more to risk prediction above and beyond what is already provided by EHR data. Although this result does not mean that integrating social determinants of health into the EHR has no benefit, researchers may be able to use EHR data alone for population risk assessment.

]]>
<![CDATA[Use of Deep Learning to Examine the Association of the Built Environment With Prevalence of Neighborhood Adult Obesity]]> https://www.researchpad.co/product?articleinfo=5c50f612d5eed0c48462bcfa

Key Points

Question

How can convolutional neural networks assist in the study of the association between the built environment and obesity prevalence?

Findings

In this cross-sectional modeling study of 4 US urban areas, extraction of built environment (ie, both natural and modified elements of the physical environment) information from images using convolutional neural networks and use of that information to assess associations between the built environment and obesity prevalence showed that physical characteristics of a neighborhood (eg, the presence of parks, highways, green streets, crosswalks, diverse housing types) can be associated with variations in obesity prevalence across different neighborhoods.

Meaning

The convolutional neural network approach allows for consistent quantification of the features of the built environment across neighborhoods and comparability across studies and geographic regions.

]]>
<![CDATA[Comparison of 2 Natural Language Processing Methods for Identification of Bleeding Among Critically Ill Patients]]> https://www.researchpad.co/product?articleinfo=5c50f5c5d5eed0c48462b75c

Key Points

Question

Can a natural language processing approach that uses text from clinical notes identify bleeding events among critically ill patients?

Findings

In this diagnostic study of a rules-based natural language processing model to identify bleeding events using clinical notes, the model was superior to a machine learning approach, with high sensitivity and negative predictive value. The extra trees machine learning model had high sensitivity but poor positive predictive value.

Meaning

Bleeding complications can be detected with a high-throughput natural language processing algorithm, an approach that can be used for quality improvement and prevention programs.

]]>
<![CDATA[Developing key performance indicators for guaranteeing right to health and access to medical service for persons with disabilities in Korea: Using a modified Delphi]]> https://www.researchpad.co/product?articleinfo=5c141ed3d5eed0c484d286ad

Recently, the Act on Guarantee of Right to Health and Access to Medical Service for Persons with Disabilities was implemented to legally define the health care system for persons with disabilities (PWDs) and the guarantee of access to medical care in Korea. This study aimed to develop specific goals and performance indicators to establish a system to guarantee right to health and access to medical service for PWDs. The first procedure was the establishment of the performance indicators, and the second was the content validity verification of the established performance indicators. To establish the performance indicators, we used the policy indicators of the government to improve the health of the Korean people. The indicators that needed to be newly developed were established based on literature review and expert consultation. Three Delphi surveys were conducted to verify the content validity of the established performance indicators. The content validity index (CVI) was obtained for the importance and possibility of the performance indicators. The indicators using the existing policy indicators are “proportion of public health centers” and “rate of health checkup of PWDs,” and newly developed indicators are “establishment of facilities for PWDs in health care facilities (buildings and personnel)” and “diagnosis of autism spectrum disorder in early childhood (average age and awareness).” The final performance indicators consist of a total of six areas, 22 sub-areas, and 40 individual indicators. The final performance indicators in this study can be used as basic data for continuously identifying the health status of PWDs in Korea and establishing the national policy for their health promotion. This study is also expected to serve as a framework to guarantee the right to health and access to medical service for PWDs rather than simply containing declarative content.

]]>
<![CDATA[Early life adiposity and telomere length across the life course: a systematic review and meta-analysis]]> https://www.researchpad.co/product?articleinfo=5c19b379d5eed0c484c53a85

Background: The relationship between adiposity at birth and in childhood, and telomere length is yet to be determined. We aimed to systematically review and meta-analyse the results of studies assessing associations between neonatal and later childhood adiposity, and telomere length.

Methods: We searched Medline, EMBASE and PubMed for studies reporting associations between adiposity measured in the neonatal period or later childhood/adolescence, and leucocyte telomere length, measured at any age via quantitative polymerase chain reaction, or terminal restriction fragment analysis, either cross-sectionally, or longitudinally. Papers published before April 2017 were included.

Results: Out of 230 abstracts assessed, 23 papers (32 estimates) were retained, from which 19 estimates were meta-analysed (15 cross-sectional, four longitudinal). Of the 15 cross-sectional estimates, seven reported on neonates: four used binary exposures of small-for-gestational-age vs. appropriate-for-gestational age (or appropriate- and large-for-gestational age), and three studied birth weight continuously. Eight estimates reported on later childhood or adolescent measures; five estimates were from studies of binary exposures (overweight/obese vs. non-obese children), and three studies used continuous measures of body mass index. All four longitudinal estimates were of neonatal adiposity, with two estimates for small-for-gestational-age vs. appropriate-for-gestational age neonates, and two estimates of birth weight studied continuously, in relation to adult telomere (49-61 years). There was no strong evidence of an association between neonatal or later childhood/adolescent adiposity, and telomere length. However, between study heterogeneity was high, and there were few combinable studies.

Conclusions: Our systematic review and meta-analysis found no strong evidence of an association between neonatal or later childhood or adolescent adiposity and telomere length.

]]>
<![CDATA[Predicting the risk of emergency admission with machine learning: Development and validation using linked electronic health records]]> https://www.researchpad.co/product?articleinfo=5bfdb37ad5eed0c4845c9d0d

Background

Emergency admissions are a major source of healthcare spending. We aimed to derive, validate, and compare conventional and machine learning models for prediction of the first emergency admission. Machine learning methods are capable of capturing complex interactions that are likely to be present when predicting less specific outcomes, such as this one.

Methods and findings

We used longitudinal data from linked electronic health records of 4.6 million patients aged 18–100 years from 389 practices across England between 1985 to 2015. The population was divided into a derivation cohort (80%, 3.75 million patients from 300 general practices) and a validation cohort (20%, 0.88 million patients from 89 general practices) from geographically distinct regions with different risk levels. We first replicated a previously reported Cox proportional hazards (CPH) model for prediction of the risk of the first emergency admission up to 24 months after baseline. This reference model was then compared with 2 machine learning models, random forest (RF) and gradient boosting classifier (GBC). The initial set of predictors for all models included 43 variables, including patient demographics, lifestyle factors, laboratory tests, currently prescribed medications, selected morbidities, and previous emergency admissions. We then added 13 more variables (marital status, prior general practice visits, and 11 additional morbidities), and also enriched all variables by incorporating temporal information whenever possible (e.g., time since first diagnosis). We also varied the prediction windows to 12, 36, 48, and 60 months after baseline and compared model performances. For internal validation, we used 5-fold cross-validation. When the initial set of variables was used, GBC outperformed RF and CPH, with an area under the receiver operating characteristic curve (AUC) of 0.779 (95% CI 0.777, 0.781), compared to 0.752 (95% CI 0.751, 0.753) and 0.740 (95% CI 0.739, 0.741), respectively. In external validation, we observed an AUC of 0.796, 0.736, and 0.736 for GBC, RF, and CPH, respectively. The addition of temporal information improved AUC across all models. In internal validation, the AUC rose to 0.848 (95% CI 0.847, 0.849), 0.825 (95% CI 0.824, 0.826), and 0.805 (95% CI 0.804, 0.806) for GBC, RF, and CPH, respectively, while the AUC in external validation rose to 0.826, 0.810, and 0.788, respectively. This enhancement also resulted in robust predictions for longer time horizons, with AUC values remaining at similar levels across all models. Overall, compared to the baseline reference CPH model, the final GBC model showed a 10.8% higher AUC (0.848 compared to 0.740) for prediction of risk of emergency admission within 24 months. GBC also showed the best calibration throughout the risk spectrum. Despite the wide range of variables included in models, our study was still limited by the number of variables included; inclusion of more variables could have further improved model performances.

Conclusions

The use of machine learning and addition of temporal information led to substantially improved discrimination and calibration for predicting the risk of emergency admission. Model performance remained stable across a range of prediction time windows and when externally validated. These findings support the potential of incorporating machine learning models into electronic health records to inform care and service planning.

]]>
<![CDATA[Disease and economic burden for rare diseases in Taiwan: A longitudinal study using Taiwan’s National Health Insurance Research Database]]> https://www.researchpad.co/product?articleinfo=5bae98da40307c0c23a1c147

Background

High-cost orphan drugs are becoming increasingly available to treat rare diseases that affect a relatively small population. Little attention has been given to the prevalence of rare diseases and their health-related economic burden in Taiwan.

Objectives

This study examined the national trends in the prevalence of rare diseases and their health-related economic burden (including medication costs) in Taiwan.

Methods

Rare disease-related claims data from 2003–2014 (12 years) from the National Health Insurance Research Database were used in this study. We used a time series analysis to assess trends in the yearly rates of treated patients with rare diseases, overall healthcare use, and expenditures, including drugs.

Results

During the 12-year study period, the estimated prevalence of rare diseases increased from 10.57 to 33.21 per 100,000 population, an average rate of a 19.46% increase per year. Total health expenditures for treatment of rare diseases increased from US$18.65 million to US$137.44 million between 2003 and 2014, accounting for 0.68% of the total national health expenditures in 2014. Drug expenditures for treatment of rare diseases increased from US$13.24 million to US$121.98 million between 2003 and 2014, which accounted for 71.00% and 88.75% of the health expenditures for patients with rare diseases in 2003 and 2014, respectively. In 2014, we found a 20.43-fold difference in average health expenditures and a 69.46-fold difference in average drug expenditures between patients with rare diseases and the overall population.

Conclusions

The prevalence of rare diseases and the related economic burden have grown substantially in Taiwan over the past 12 years, and these trends are likely to continue. Drug expenditures accounted for almost 90% of health expenditures for rare diseases. Further analyses are underway to examine the economic burden of individual rare diseases.

]]>
<![CDATA[<i>PLoS Medicine</i> Issue Image | Vol. 15(11) November 2018]]> https://www.researchpad.co/product?articleinfo=5c0ae483d5eed0c484589dd7

Advancing the beneficial use of machine learning in health care and medicine: Toward a community understanding

Modern statistical modeling techniques—often called machine learning—are posited as a transformative force for human health. This month we present our Special Issue, Guest Edited by Atul Butte, Suchi Saria, and Aziz Sheikh, on machine learning-based advances in clinical care, health systems, and pathophysiology.

The Special Issue provides a broad perspective on what machine learning can currently achieve in health and biomedicine, and presents expert commentary on future benefits and constraints of bringing machine learning into clinical care, population health, and public policy.

In our accompanying Editorial, the PLOS Medicine Editors parlay the lessons we've learned during the development of the Special Issue into guidance for generating machine learning-based models that are demonstrably valid and fit-for-purpose.

Image Credit: StockSnap, Pixabay

]]>
<![CDATA[Rival perspectives in health technology assessment and other economic evaluations for investing in global and national health. Who decides? Who pays?]]> https://www.researchpad.co/product?articleinfo=5c02f1edd5eed0c4844a16e7

There seems to be a general agreement amongst practitioners of economic evaluations, including Health Technology Assessment, that the explicit statement of a perspective is a necessary element in designing and reporting research. Moreover, there seems also to be a general presumption that the ideal perspective is “societal”. In this paper we endorse the first principle but dissent from the second. A review of recommended perspectives is presented. The societal perspective is frequently not the one recommended. The societal perspective is shown to be less comprehensive than is commonly supposed, is inappropriate in many contexts and, in any case, is in general not a perspective to be determined independently of the context of a decision problem. Moreover, the selection of a perspective, societal or otherwise, is not the prerogative of analysts.

]]>
<![CDATA[Measuring colorectal cancer incidence: the performance of an algorithm using administrative health data]]> https://www.researchpad.co/product?articleinfo=5b5989b3463d7e76cf8ed8b6

Background

Certain cancer case ascertainment methods used in Quebec and elsewhere are known to underestimate the burden of cancer, particularly for some subgroups. Algorithms using claims data are a low-cost option to improve the quality of cancer surveillance, but have not frequently been implemented at the population-level. Our objectives were to 1) develop a colorectal cancer (CRC) case ascertainment algorithm using population-level hospitalization and physician billing data, 2) validate the algorithm, and 3) describe the characteristics of cases.

Methods

We linked physician billing, hospitalization, and tumor registry data for 2,013,430 Montreal residents age 20+ (2000–2010). We compared the performance of three algorithms based on diagnosis and treatment codes from different data sources. We described identified cases according to age, sex, socioeconomic status, treatment patterns, site distribution, and time trends. All statistical tests were two-sided.

Results

Our algorithm based on diagnosis and treatment codes identified 11,476 of the 12,933 incident CRC cases contained in the tumor registry as well as 2317 newly-captured cases. Our cases share similar overall time trends and site distributions to existing data, which increases our confidence in the algorithm. Our algorithm captured proportionally 35% more individuals age 50 and younger among CRC cases: 8.2% vs. 5.3%. The newly captured cases were also more likely to be living in socioeconomically advantaged areas.

Conclusions

Our algorithm provides a more complete picture of population-wide CRC incidence than existing case ascertainment methods. It could be used to estimate long-term incidence trends, aid in timely surveillance, and to inform interventions, in both Quebec and other jurisdictions.

Electronic supplementary material

The online version of this article (10.1186/s12874-018-0494-x) contains supplementary material, which is available to authorized users.

]]>
<![CDATA[Is there a place for undergraduate and graduate students in the systematic review process?]]> https://www.researchpad.co/product?articleinfo=5b4d0bf5463d7e142d5b64e4

Systematic reviews are a well-established and well-honed research methodology in the medical and health sciences fields. As the popularity of systematic reviews has increased, disciplines outside the sciences have started publishing them. This increase in familiarity has begun to trickle down from practitioners and faculty to graduate students and recently undergraduates. The amount of work and rigor that goes into producing a quality systematic review may make these types of research projects seem unattainable for undergraduate or graduate students, but is this an accurate assumption? This commentary discusses whether there is a place for undergraduate and graduate students in the systematic review process. It explains the possible benefits of having undergraduate and graduate students engage in systematic reviews and concludes with ideas for creating basic education or training opportunities for researchers and students who are new to the systematic review process.

]]>