Explore the vast range of titles available.

The majority of people with hypertension require more than one medication to achieve blood pressure control. Many patients are prescribed multipill antihypertensive regimens rather than single-pill fixed-dose combination (FDC) treatment. Although FDC use may improve medication adherence, the impact on patient outcomes is unclear. We compared clinical outcomes and medication adherence with FDC therapy versus multipill combination therapy in a real-world setting using linked clinical and administrative databases. We conducted a population-based retrospective cohort study of 13,350 individuals 66 years and older in Ontario, Canada with up to 5 years of follow-up. We included individuals who were newly initiated on one angiotensin-converting enzyme inhibitor (ACEI) or angiotensin II-receptor blocker (ARB) plus one thiazide diuretic. High-dimensional propensity score matching was used to compare individuals receiving FDC versus multipill therapy. The primary outcome was a composite of death or hospitalization for acute myocardial infarction (AMI), heart failure, or stroke. We conducted 2 analyses to examine the association between adherence and patient outcomes. First, we performed an on-treatment analysis to determine whether outcomes differed between groups while patients were on treatment, censoring patients when they first discontinued treatment, defined as not receiving medications within 150% of the previous days' supply. Second, we conducted an intention-to-treat analysis that followed individuals allowing for breaks in treatment to quantify the difference in drug adherence between groups and assess its impact on clinical outcomes. As expected, there was no significant difference in the primary outcome between groups in the on-treatment analysis (HR 1.06, 95% CI 0.86-1.31, P = 0.60). In the intention-to-treat analysis, the proportion of total follow-up days covered with medications was significantly greater in the FDC group (70%; IQR 19-98) than in the multipill group (42%, IQR 11-91, P < 0.01), and the primary outcome was less frequent in FDC recipients (3.4 versus 3.9 events per 100 person-years; HR 0.89, 95% CI 0.81-0.97, P < 0.01). The main limitations of this study were the lack of data regarding cause of death and blood pressure measurements and the possibility of residual confounding. Among older adults initiating combination antihypertensive treatment, FDC therapy was associated with a significantly lower risk of composite clinical outcomes, which may be related to better medication adherence.

Prescription opioid use is highly associated with risk of opioid-related death, with 1 of every 550 chronic opioid users dying within approximately 2.5 years of their first opioid prescription. Although gabapentin is widely perceived as safe, drug-induced respiratory depression has been described when gabapentin is used alone or in combination with other medications. Because gabapentin and opioids are both commonly prescribed for pain, the likelihood of co-prescription is high. However, no published studies have examined whether concomitant gabapentin therapy is associated with an increased risk of accidental opioid-related death in patients receiving opioids. The objective of this study was to investigate whether co-prescription of opioids and gabapentin is associated with an increased risk of accidental opioid-related mortality. We conducted a population-based nested case-control study among opioid users who were residents of Ontario, Canada, between August 1, 1997, and December 31, 2013, using administrative databases. Cases, defined as opioid users who died of an opioid-related cause, were matched with up to 4 controls who also used opioids on age, sex, year of index date, history of chronic kidney disease, and a disease risk index. After matching, we included 1,256 cases and 4,619 controls. The primary exposure was concomitant gabapentin use in the 120 days preceding the index date. A secondary analysis characterized gabapentin dose as low (<900 mg daily), moderate (900 to 1,799 mg daily), or high (≥1,800 mg daily). A sensitivity analysis examined the effect of concomitant nonsteroidal anti-inflammatory drug (NSAID) use in the preceding 120 days. Overall, 12.3% of cases (155 of 1,256) and 6.8% of controls (313 of 4,619) were prescribed gabapentin in the prior 120 days. After multivariable adjustment, co-prescription of opioids and gabapentin was associated with a significantly increased odds of opioid-related death (odds ratio [OR] 1.99, 95% CI 1.61 to 2.47, p < 0.001; adjusted OR [aOR] 1.49, 95% CI 1.18 to 1.88, p < 0.001) compared to opioid prescription alone. In the dose-response analysis, moderate-dose (OR 2.05, 95% CI 1.46 to 2.87, p < 0.001; aOR 1.56, 95% CI 1.06 to 2.28, p = 0.024) and high-dose (OR 2.20, 95% CI 1.58 to 3.08, p < 0.001; aOR 1.58, 95% CI 1.09 to 2.27, p = 0.015) gabapentin use was associated with a nearly 60% increase in the odds of opioid-related death relative to no concomitant gabapentin use. As expected, we found no significant association between co-prescription of opioids and NSAIDs and opioid-related death (OR 1.11, 95% CI 0.98 to 1.27, p = 0.113; aOR 1.14, 95% CI 0.98 to 1.32, p = 0.083). In an exploratory analysis of patients at risk of combined opioid and gabapentin use, we found that 46.0% (45,173 of 98,288) of gabapentin users in calendar year 2013 received at least 1 concomitant prescription for an opioid. This study was limited to individuals eligible for public drug coverage in Ontario, we were only able to identify prescriptions reimbursed by the government and dispensed from retail pharmacies, and information on indication for gabapentin use was not available. Furthermore, as with all observational studies, confounding due to unmeasured variables is a potential source of bias. In this study we found that among patients receiving prescription opioids, concomitant treatment with gabapentin was associated with a substantial increase in the risk of opioid-related death. Clinicians should consider carefully whether to continue prescribing this combination of products and, when the combination is deemed necessary, should closely monitor their patients and adjust opioid dose accordingly. Future research should investigate whether a similar interaction exists between pregabalin and opioids.

The provision of health care is hard and, understandably, clinicians struggle. Beyond meaningful health care interactions, clinicians must also integrate large volumes of demographic, diagnostic, and prognostic data with the latest clinical evidence and guidelines to arrive at a diagnosis, determine optimal treatment pathways based on prognosis, communicate clearly with patients, and help patients navigate services -- all in stressful, time-constrained, uncertain situations. Asthma is misdiagnosed in about one-third of patients. Physicians correctly predict in-hospital death less than one-third of the time. Treatment pathways for many conditions must necessarily follow a trial-and-error approach, given substantial variability in treatment responses and adverse events. At the health system level, hospitals are often over capacity, health care professionals are stretched, wait times for specialist appointments and elective surgical procedures are relatively long, and patients are only moderately satisfied with the care they receive. Clinicians and health system leaders need all the help they can get, and everyone knows it.

The use of opioids for noncancer pain is widespread, and more than 16,000 die of opioid-related causes in the United States annually. The patients at greatest risk of death are those receiving high doses of opioids. Whether sex influences the risk of dose escalation or opioid-related mortality is unknown. We conducted a cohort study using healthcare records of 32,499 individuals aged 15 to 64 who commenced chronic opioid therapy for noncancer pain between April 1, 1997 and December 31, 2010 in Ontario, Canada. Patients were followed from their first opioid prescription until discontinuation of therapy, death from any cause or the end of the study period. Among patients receiving chronic opioid therapy, 589 (1.8%) escalated to high dose therapy and n = 59 (0.2%) died of opioid-related causes while on treatment. After multivariable adjustment, men were more likely than women to escalate to high-dose opioid therapy (adjusted hazard ratio 1.44; 95% confidence interval 1.21 to 1.70) and twice as likely to die of opioid-related causes (adjusted hazard ratio 2.04; 95% confidence interval 1.18 to 3.53). These associations were maintained in a secondary analysis of 285,520 individuals receiving any opioid regardless of the duration of therapy. Men are at higher risk than women for escalation to high-dose opioid therapy and death from opioid-related causes. Both outcomes were more common than anticipated.

The spectrum of clinical settings in which machine learning approaches have been examined for use in the health care setting has increased markedly and become more diverse in recent years. Many studies have detailed the data science and statistical bases of machine-learned tools. However, comparatively few studies have focused on their evaluation and implementation. Here, Antoniou discusses how to evaluate machine-learned solutions throughout their life cycle to optimize their use and functionality in clinical practice.

Machine learning (ML) holds great promise in transforming healthcare. While published studies have shown the utility of ML models in interpreting medical imaging examinations, these are often evaluated under laboratory settings. The importance of real world evaluation is best illustrated by case studies that have documented successes and failures in the translation of these models into clinical environments. A key prerequisite for the clinical adoption of these technologies is demonstrating generalizable ML model performance under real world circumstances. The purpose of this study was to demonstrate that ML model generalizability is achievable in medical imaging with the detection of intracranial hemorrhage (ICH) on non-contrast computed tomography (CT) scans serving as the use case. An ML model was trained using 21,784 scans from the RSNA Intracranial Hemorrhage CT dataset while generalizability was evaluated using an external validation dataset obtained from our busy trauma and neurosurgical center. This real world external validation dataset consisted of every unenhanced head CT scan (n = 5965) performed in our emergency department in 2019 without exclusion. The model demonstrated an AUC of 98.4%, sensitivity of 98.8%, and specificity of 98.0%, on the test dataset. On external validation, the model demonstrated an AUC of 95.4%, sensitivity of 91.3%, and specificity of 94.1%. Evaluating the ML model using a real world external validation dataset that is temporally and geographically distinct from the training dataset indicates that ML generalizability is achievable in medical imaging applications.

Opioid prescribing and overdose are leading public health problems in North America, yet the precise public health burden has not been quantified. To examine the burden of opioid-related mortality across the United States over time. This study used a serial cross-sectional design in which cross sections were examined at different time points to investigate deaths from opioid-related causes in the United States between January 1, 2001, and December 31, 2016. Opioid-related deaths, defined as those in which a prescription or illicit opioid contributed substantially to an individual's cause of death as determined by death certificates. We compared the percentage of deaths attributable to opioids and the associated person-years of life lost by age group. Between 2001 and 2016, the number of opioid-related deaths in the United States increased by 345%, from 9489 to 42 245 deaths (33.3 to 130.7 deaths per million population). By 2016, men accounted for 67.5% of all opioid-related deaths, and the median (interquartile range) age at death was 40 (30-52) years. The percentage of deaths attributable to opioids increased in a similar fashion. In 2001, 0.4% of deaths (1 in 255) were opioid related, rising to 1.5% of deaths (1 in 65) by 2016, an increase of 292%. This burden was highest among adults aged 24 to 35 years. In this age group, 20.0% of deaths were attributable to opioids in 2016. Among those aged 15 to 24 years, 12.4% of deaths were attributable to opioids in 2016. Overall, opioid-related deaths resulted in 1 681 359 years of life lost (5.2 per 1000 population) in the United States in 2016, most of which (1 125 711 years of life lost) were among men. Adults aged 25 to 34 years had 12.9 years of life lost per 1000 population, and those aged 35 to 44 years had 9.9 years of life lost per 1000 population. Premature death from opioid-related causes imposes an enormous public health burden across the United States. The recent increase in deaths attributable to opioids among those aged 15 to 34 years highlights a need for targeted programs and policies that focus on improved addiction care and harm reduction measures in this high-risk population.

The recent publication of a national guideline and quality standards in Canada have provided clinicians with new, evidence-based recommendations on safe, appropriate opioid use. We sought to characterize how well opioid initiation practices aligned with these recommendations before and following their release. We conducted a population-based study among people initiating opioids prior to the release of national guidelines (April 2015-March 2016; fiscal year [FY] 2015) and in the most recent year available (January-December 2019) in Ontario, Canada. We used linked administrative claims data to ascertain the apparent indication for opioid therapy, and characterized the initial daily dose (milligrams morphine or equivalent; MME) and prescription duration for each indication. In FY2015, 653,885 individuals commenced opioids, compared to 571,652 in 2019. Over time, there were small overall reductions in the prevalence of initial daily doses exceeding 50MME (23.9% vs. 20.1%) and durations exceeding 7 days (17.4% vs. 14.8%); but the magnitude of the reductions varied widely by indication. The prevalence of high dose (>50MME) initial prescriptions reduced significantly across all indications, with the exception of dentist-prescribed opioids (13.6% vs. 12.1% above 50MME). In contrast, there was little change in initial durations exceeding 7 days across most indications, with the exception of some surgical indications (e.g. common excision; 9.3% vs. 6.2%) and among those in palliative care (35.2% vs. 29.2%). Despite some modest reductions in initiation of high dose and long duration prescription opioids between 2015 and 2019, clinical practice is highly variable, with opioid prescribing practices influenced by clinical indication. These findings may help identify medical specialties well-suited to targeted interventions to promote safer opioid prescribing.

Machine learning (ML) is a rapidly advancing field with increasing utility in health care. We conducted a systematic review and critical appraisal of ML applications in vascular surgery. MEDLINE, Embase, and Cochrane CENTRAL were searched from inception to March 1, 2021. Study screening, data extraction, and quality assessment were performed by two independent reviewers, with a third author resolving discrepancies. All original studies reporting ML applications in vascular surgery were included. Publication trends, disease conditions, methodologies, and outcomes were summarized. Critical appraisal was conducted using the PROBAST risk-of-bias and TRIPOD reporting adherence tools. We included 212 studies from a pool of 2235 unique articles. ML techniques were used for diagnosis, prognosis, and image segmentation in carotid stenosis, aortic aneurysm/dissection, peripheral artery disease, diabetic foot ulcer, venous disease, and renal artery stenosis. The number of publications on ML in vascular surgery increased from 1 (1991–1996) to 118 (2016–2021). Most studies were retrospective and single center, with no randomized controlled trials. The median area under the receiver operating characteristic curve (AUROC) was 0.88 (range 0.61–1.00), with 79.5% [62/78] studies reporting AUROC ≥ 0.80. Out of 22 studies comparing ML techniques to existing prediction tools, clinicians, or traditional regression models, 20 performed better and 2 performed similarly. Overall, 94.8% (201/212) studies had high risk-of-bias and adherence to reporting standards was poor with a rate of 41.4%. Despite improvements over time, study quality and reporting remain inadequate. Future studies should consider standardized tools such as PROBAST and TRIPOD to improve study quality and clinical applicability.