Artificial Intelligence in a Smartwatch for Atrial Fibrillation

An interesting study by Perez et al in NEJM found some very interesting things:

1. A large number of people opted-in to be part of the study (over 400K)
2. In about 4 months, about 0.5% participants were noted to have irregular heart rate – a relatively small percentage than one would expect.
3. Of the people who returned 7-day monitoring after being notified of irregular heart beat, about one-third had atrial fibrillation. This is a significant number of atrial fibrillation in otherwise asymptomatic individuals.

The study highlights several important things, some noted above. However, what it does not tell us if the use of smartwatch to identify atrial fibrillation in otherwise healthy people results in improved health outcomes or not. A future study should be able to evaluate this question. For now, at least we know we can depend on smartwatch to identify atrial fibrillation.

Burnout among Nurses

Dyrbye et al conducted a national survey of the US nurses (response rate 26.2%) and found that 35.3% had burnout symptoms. Interestingly, 30.7% of the respondents also had symptoms of depression. Quite interestingly, authors found that nurses who had higher burnout were more likely to have been absent 1 or more days during the last month and had poor work-performance.

Critique: Authors need to be commended for conducting a national study of nurses. However, this study has the same issue that we find in other burnout literature; person who is burnout is also the one who determines if they have poor performance or not. At least theoretically, one would assume that a person who is burnout will look more negatively towards themselves and their work than a person who is not burned out; we would expect a high correlation between the two. On the other hand, a study in which nurses’ superiors could have been asked to comment on performance would have generated more reliable data, however, such study would have been quite expensive requiring significant external funding.

Burnout in Healthcare–A systems approach

Interesting and thought-provoking analytical piece by Montgomery et al which correctly points out that the responsibility of mitigating burnout lies with the institution and that the practice of putting the responsibility on healthcare workers should be abandoned. Currently, most places try to focus on the healthcare worker – asking them to take better care of themselves, be resilient, cope with stressors on their own, etc. They propose the following:

1. Provider burnout should be added to the current assessment of healthcare quality
2. Burnout should be assessed at the departmental/unit level (and I would add that it should be measured at least annually).
3. Development of health workplaces should be a major goal
4. Key questions concerning research and practice should not be coming from researchers alone.

Personally, I do believe that there are not enough researchers examining the interventions needed to address burnout. There is lack of funding from national funding agencies and from locally from institutions.

Some Recent Physician Burnout Literature

An abstract, presented at the the annual meeting and published in the Journal of American College of Surgeons’ supplement examined relationship between burnout and involvement of surgeons in medical device development/innovation.

As is often found in medical students across the world, New Zealand’s medical students also appear to have high levels of burnout. Farrell et al surveyed one of the two medical school in New Zealand and found that the prevalence of exhaustion was 77% while of disengagement was 68%. Another proof that burnout in medical profession is a worldwide phenomenon.

A study examined burnout among Egyptian ICU physicians and found interesting results. The level of exhaustion was 38.8%, somewhat better than generally found in the West. However, the level of depersonalization was 70.1%. This is really surprising; do Egyptian physicians really feel that much disconnected or disengaged from their patients?

Should Older Patients with Subclinical Hypothyroidism be Treated with Thyroid Replacement?

With increasing age, many patients develop subclinical hypothyroidism. Subclinical hypothyroidism is defined by the presence of elevated thyroid stimulating hormone (TSH) but normal free T4 levels. Some patients with subclinical hypothyroidism also have symptoms that are not uncommon with increasing age such as constipation, tiredness, mental slowness, and fatigue.

Prior studies have found no association between treating patients with subclinical hypothyroidism with thyroid replacement therapy and resolution of symptoms. However, very old individuals (older than 80 years) are often not included in the studies.

To examine if treatment of subclinical hypothyroidism with thyroid replacement therapy results in resolution of such symptoms, Moojiraat et al. combined data from two randomized controlled trials. Combining data from two trials results in increased number of patients above 80 (and hence power of the study to detect a difference).

Contrary to expectations, but consistent with previous findings, thyroid replacement therapy does not result in resolution of symptoms in patients with subclinical hypothyroidism. Instead, there may be some increase in adverse events.

Prediction of COPD Exacerbations – ACCEPT Tool

Preventing exacerbations in patients with COPD is a major goal. If we can identify patients who are at higher risk of exacerbations, we may be able to tailor more aggressive therapy to such patients. This will result in better utilization of resources, improved risk/benefit ratio, and will not expose low-risk patients to the adverse effects of aggressive therapies. Thus, predicting exacerbation risk in individual patients can guide these clinical decisions. Unfortunately, there are no externally validated and implementable tools to predict COPD exacerbation.

That is until now: Adibi and colleagues have uploaded their manuscript to a preprint server (bioRxiv) which used data from three randomized trials to develop ACCEPT, a clinical prediction tool based on routinely available predictors for COPD exacerbations. Authors externally validated ACCEPT in a large, multinational prospective cohort. ACCEPT appears to be the first COPD exacerbation prediction tool that jointly estimates the individualized rate and severity of exacerbations. The tool is designed to be easily applicable in clinical practice and is accessible as a web application.

Issues with Current Physician Burnout Research

The relationship of physician burnout with patient outcomes has been often reported but the studies have been of poor quality and/or report the outcomes as perceived by physicians. Obviously, using physician’s perception of burnout and physician’s perception of patient outcomes is circular in nature; a burnout physician may be likely to think that the patient care provided by her is of lower quality.

A systemic review, in Annals of Internal Medicine highlights the similar issue. Moreover, it also shows that where a relationship is shown, the effect sizes tend to be larger suggesting that publication bias may be contributing to preponderance of studies showing an effect (because studies showing no effect are not getting published.

The physician burnout research (and researchers) should start using rigorous scientific methods to define outcomes, predictors, associations, and evaluation of interventions. Otherwise, such an important topic will get a poor reputation due to low-quality research.

National Academy of Medicine Talks about Physician Burnout

In response to concerning rates of depression, stress, and burnout among US medical students and clinicians, the National Academy of Medicine (formerly Institute of Medicine) launched the Action Collaborative on Clinician Well-Being and Resilience in 2017.

Incorporating input from experts in human factors and systems engineering and health informatics, as well as medical, nursing, pharmacy, and dentistry experts and educators, the committee took a systems approach to clinician burnout.

The committee’s systems model for professional well-being and clinician burnout has 3 levels: frontline care delivery, health care organization, and external environment, which together influence the work system factors that contribute to clinician burnout and professional well-being. The work system factors often extend across more than 1 system level (care delivery, health care organization, and external environment), and improvement can occur at every level to relieve workplace stress. The recommendations from the report, organized under 6 overarching goals, reflect the crosscutting nature of the identified factors contributing to clinician burnout and professional well-being.

1. Create positive work environments

2. Create positive learning environments

3. Reduce administrative burden

4. Enable technology solutions

5. Provide support to clinicians and students

6. Invest in research

Angiotensin Receptor Blockers and Suicide Risk

Interesting conclusions in a study reported in the JAMA Network Open:

The use of ARBs may be associated with an increased risk of suicide compared with ACEIs.

Investigators matched 964 cases to 3856 controls. Compared to ACE inhibitors, ARBs were associated with 63% higher risk of death by suicide. The results remained significant (60% increase) when individuals with history of self-harm were excluded.

Proliferation of Risk Factors for Physician Burnout

Several authors have promoted various risk factors for physician burnout. Below are some that I noticed in my email:

Moral Injury:

Kopacz MS, Ames D, Koenig HG. It's time to talk about physician burnout and moral injury. Lancet Psychiatry. 2019 Nov;6(11):e28. doi:10.1016/S2215-0366(19)30385-2. PubMed PMID: 31631880.

Low or lack of Mindfulness:

Lebares CC et al. Key factors for implementing mindfulness-based burnout interventions in surgery. The American Journal of Surgery (in Press)

Lack of Resilience and Grit

Shakir HJ, Cappuzzo JM, Shallwani H, Kwasnicki A, Bullis C, Wang J, Hess RM,Levy EI. Relationship of Grit and Resilience to Burnout Among US NeurosurgeryResidents. World Neurosurg. 2019 Oct 16. pii: S1878- 750(19)32658-0. doi:10.1016/j.wneu.2019.10.043. [Epub ahead of print]  PubMed PMID: 31629138.

Attacks on the Calling of Medicine

Stewart MT, Serwint JR. Burning without burning out: A call to protect thecalling of medicine. Curr Probl Pediatr Adolesc Health Care. 2019 Oct 17:100655. doi: 10.1016/j.cppeds.2019.100655. [Epub ahead of print] PubMed PMID: 31631025

Above are some of the examples of risk factors cited in literature and the list continues to grow with each passing day. There are folks who have developed ‘Burnout Prevention Programs’ around these risk factors. Some even market these programs and make money.

My concern stems from the fact that while there may be strong conceptual underpinnings for at least some of these risk factors, rigorous studies examining the relationship are missing. The current state of evidence is so poor that a clinician will not take seriously risk factors with such limited evidence when managing a patient. Then, why, physicians doing this to themselves.

In particular, the worrisome part is the so-called prevention programs. Where is the evidence to support that such interventions work? If we don’t accept evidence for our patients without randomized clinical trials or at least strong observational studies, why we accept such intervention programs without evidence of efficacy?

Waste in US Healthcare System

US healthcare system is quite unique; it is the costliest system in the world but the outcomes are poor. In other words, the value of healthcare (outcomes/cost) is low and a significant amount expense goes in waste. There have been efforts to reduce waste in the healthcare system but it is unclear how effective such efforts have been.

In one estimate, the overall annual cost of waste in the healthcare is between $760 to $935 billion or 25% of the total healthcare spending. Authors also tried to estimate the amount of waste in different domains. Of interest is that fact that administrative complexity and pricing failure are the largest ticket items but there are only meager attempts to address these.

Below are the numbers:

Domain	Waste (in billions)
Failure of Care Delivery	$102.4 to $165.7
Failure of Care Coordination	$27.2 to $78.2
Low-value Care	$75.7 to $101.2
Pricing Failure	$230.7 to $240.5
Fraud and Abuse	$58.5 to $83.9
Administrative Complexity	$265.6

Selepressin: Not so useful in Septic Shock

Septic shock is an advanced stage of body’s response to an infection and manifests as marked decrease in blood pressure with resulting decrease blood flow (and hence nutrients and oxygen) to the tissues. The decreased blood pressure is due to vasodilatation and increased capillary permeability (leaky capillaries).

The treatment of septic shock includes antibiotics to treat the infection, intravenous fluids to replenish fluid that has seeped out into the tissues from leaky capillaries, and vasopressors (such as norepinephrine) to counter vasodilation. Not uncommonly, norepinephrine is not sufficient to raise blood pressure and increase blood flow to the tissues. Vasopressin, another vasopressor, is sometimes used in addition to norepinephrine to support blood pressure. However, vasopressin has other adverse effects. These adverse effects are due to the fact that vasopressin stimulates three types of vasopressin receptors (V1a, V1b, and V2). Stimulation of V1a has vasopressor effect while stimulation of V1b and V2 results in increased coagulation activity, nitric oxide release, corticosteroid secretion, and excessive water retention.

Selepressin is a selective V1a receptor agonist and has only vasopressor effect. One would assume that selective stimulation of V1a receptors with selepressin will result in beneficial outcome. However, in a clinical trial (N=868), selepressin was not found to be effective in reducing ventilator-free days, norepinephrine-free days, mortality, lower ICU days, or lower need for kidney replacement therapy.

High-Flow Nasal Cannula vs. Noninvasive Ventilation

In patients who are being mechanically ventilated and are at high risk of extubation failure, this randomized controlled trial (N=641) found non-invasive ventilation with high-flow oxygen superior to high-flow nasal cannula alone. Here is the study.

Hospital Readmission Penalty Might be Increasing Mortality

Hospital Readmission Reduction Program and association penalties for higher than average/expected readmissions have resulted in a significant decrease in 30-day readmissions after hospital discharge. This also appears to have saved money to the CMS. When looking at the readmissions only, this programs appears to be a resounding success. However, its unintended consequences are becoming clear only now.

Using a national database of almost all hospitals, this study found that hospitals that were able to decrease readmission rates for patients with acute exacerbation of COPD, also had an increase in mortality for such patients within 30-days after discharge. While the underlying mechanisms are open for speculation, this association needs to be taken seriously and possibility of a casual relationship needs to be explored.

Stacked ICU Admissions and Mortality

This interesting study shows that when ICU admissions are stacked, that is two or more admissions come too close to each other, there is an increased risk of patient mortality, longer hospital stay, and higher odds of nursing home discharge.

Investigators enrolled 13,234 consecutive ICU admissions of which 1/4rth had an elapsed time since the last admission (ETLA) of < 55 min. Stacked admissions had on average, a higher unadjusted [1.16 (95% CI 1–1.35, P = 0.05)] and adjusted [1.23 (95% CI 1.04–1.44, P = 0.01)] odds ratio of ICU death, higher unadjusted [1.11 (95% CI 0.99–1.24, P = 0.06)] and adjusted [1.20 (95% 1.05–1.35, P = 0.004)]  odds ratio of hospital death, and a lower adjusted OR of home discharge of 0.91 (95% CI 0.84–0.99, P = 0.04).

Log-transformed Predictor in Regression Model

It is not uncommon in a regression model that a predictor is log-transformed to meet the normality assumption of the residuals. Below is an example where our goal is to examine a relationship between urinary arsenic concentration and white blood cell (WBC) count (in thousands). Urinary arsenic distribution had right-skew and hence the predictor was log-transformed for this regression. The output is below and the coefficient is highlighted in yellow.

The interpretation of regression coefficients can be sometimes confusing. However, when the predictor variable is a continuous variable (here it is LNUARS), it is easy to visualize it graphically. Simply, think that the coefficient is slope for a line on a graph where Y-axis has outcome (WBC count here) and X-axis has predictor. Now, we can interpret it as ‘change in Y (WBC here) for each unit change in X (LNUARS here)’. Note, we are saying a unit change and this unit can be any unit depending on a given variable.

Now, we have our predictor (urinary arsenic) log-transformed due to its skewed distribution. The coefficient (or slope of the graph) here means change in WBC count (unit is in thousands for this output) for one unit change in log of total normalized urinary arsenic. While this is an accurate interpretation of the coefficient, we don’t use log-scale measurements in our regular life. Further, we may find it difficult to communicate with others when describing results. Hence, it makes much more sense to convert total arsenic from log-scale to our usual scale.

As a general rule, and without going into mathematical details, the interpretation of a log-transformed variable is slightly different than usual interpretation that we would do otherwise. A simplest way is to multiply the coefficient with 0.01; the resulting value will be change in Y for 1% change in X. Note, it is not one unit change bur rather one percent change. Here, the coefficient is -0.195. Multiplying it with 0.01 gives us -0.00195. The Y = WBC here has unit in 1000 cells and X here is total urinary arsenic. Hence, we will say that for each 1% increase in total urinary arsenic, the WBC decreases by 0.00195 (in thousands). We can multiply 0.00195 by 1000 (=1.95) and then each 1% increase in normalized total urinary arsenic decreases WBC by about 2 cells.

The p-value is significant but the change of 2-cells for 1% change in urinary arsenic may not be large enough to be clinically meaningful; however, that is another topic of discussion – difference between statistically significant and clinically meaningful – for another day.

Anti-Mullerian Hormone in Men

Anti-Müllerian hormone (AMH) is a Sertoli cell-secreted protein that plays a major role in the development of internal male genitalia during embryonic life. Around the 7th week of gestation, AMH causes regression of the Mullerian duct and hence it is also known as Müllerian-inhibiting substance (MIS). Persistent Mullerian duct leads to formation of female internal sex organs. During adult life, AMH continues to be produced by the Sertoli cells in the testis in men although its functional relevance remains unclear.

In 2016, an very strong association of AMH with all-cause mortality was reported in men.

“In unadjusted analysis, each unit increase in serum anti-mullerian hormone level was associated with a 13 % lower risk of death (HR = 0.87; 95 %CI = 0.83-0.92). In multivariable models, the inverse association between serum anti-mullerian hormone levels and mortality remained significant (HR = 0.94; 95 %CI = 0.90-0.98) and was independent of confounding variables. Similarly, individuals in the highest quartile had significantly lower risk of death as compared to individuals in the lowest quartile (unadjusted HR = 0.13, 95 %CI = 0.07-0.25; adjusted HR = 0.36, 95 %CI = 0.16-0.81).”

While the study showed an association the underlying mechanistic pathways remained unclear.

Recently, AMH has been shown to be associated with serum C-reactive protein (CRP) levels in men raising the possibility that the underlying mechanism may include modulation of inflammatory response. It is a potentially an exciting area of research and new discoveries in future may highlight important relationships between AMH and health, morbidity, and mortality in humans.

Soft Drink Consumption and Mortality

In this population-based cohort study of 451,743 individuals from 10 countries in Europe, greater consumption of total, sugar-sweetened, and artificially sweetened soft drinks was associated with a higher risk of all-cause mortality.

1. 17% higher all-cause mortality was found among participants who consumed 2 or more glasses per day (vs consumers of <1 glass per month) of total soft drinks (hazard ratio [HR], 1.17; 95% CI, 1.11-1.22; P < .001),

2. 8% higher mortality in participants who consumed sugar-sweetened soft drinks (HR, 1.08; 95% CI, 1.01-1.16; P = .004), and

3. 26% higher mortality in participants who drank artificially sweetened soft drinks (HR, 1.26; 95% CI, 1.16-1.35; P < .001).

Consumption of artificially sweetened soft drinks was positively associated with deaths from circulatory diseases, and sugar-sweetened soft drinks were associated with deaths from digestive diseases.

Direct Oral Anticoagulants vs Warfarin in Older Patients With Atrial Fibrillation and Ischemic Stroke

This observational study examined 11,662 patients with atrial fibrillation who had had an ischemic stroke and were anticoagulation naïve, patients discharged while receiving:

1. Direct oral anticoagulants (DOACs) had more days at home post-discharge
2. Were less likely to experience major adverse cardiovascular events, all-cause mortality, all-cause readmissions, cardiovascular readmissions, or hemorrhagic strokes
3. Had a small but significant increase in gastrointestinal bleeding.

Overall, DOACs appear to be an effective and safe treatment option compared with warfarin for patients with atrial fibrillation who have ischemic stroke and may be even more beneficial despite a small in crease in gastrointestinal bleeding.

Perioperative Direct Oral Anticoagulant Use

This study of 3007 patients with atrial fibrillation who were undergoing surgery enrolled patients to have:

1. Direct Oral Anticoagulants (DOACs) were omitted 1-day before a low-bleeding-risk procedure
2. DOACs were omitted 2 days before a high-bleeding-risk procedure
3. Patients with renal impairment had the pre-operative omission duration readjusted

Study investigators found that the perioperative risk of bleeding and arterial thromboembolism were low with this simplified strategy. The 30-day postoperative rate of major bleeding was 1.35% (95% CI, 0%-2.00%) in the apixaban cohort, 0.90% (95% CI, 0%-1.73%) in the dabigatran cohort, and 1.85% (95% CI, 0%-2.65%) in the rivaroxaban cohort. The rate of arterial thromboembolism was 0.16% (95% CI, 0%-0.48%) in the apixaban cohort, 0.60% (95% CI, 0%-1.33%) in the dabigatran cohort, and 0.37% (95% CI, 0%-0.82%) in the rivaroxaban cohort.

DOACs	Bleeding Risk	Arterial Thromboembolism
Apixaban	1.35	0.16
Dabigatran	0.90	0.60
Rivaroxaban	1.85	0.37