Study shows more work, fewer new patients, little health benefit
Telemedicine and other "e-visits" are supposed to be a win-win for physicians and patients alike. Doctors could spend less time on simple requests, patients would get frictionless access to their provider. But a new study published in Management Science finds that all that access hasn't translated into the outcomes so many had hoped for. Instead, e-visits lead to more office visits and more phone consultations without measurable improvement to patients' health. And maybe most damaging for physicians' practices, they're associated with fewer new patients.
The findings may be surprising, but study leader Hessam Bavafa, PhD, of the University of Wisconsin School of Business, said they make sense when you consider the process of the usual e-visit. Patients can reach out with even the smallest concerns, he said, and that puts doctors in a bind." There's an issue of obligation," Bavafa said. "If you ignore the signal, who knows what's going to happen next, right?"
The study used five years of data from a large health system with multiple hospitals and more than 2,000 total beds. It included all primary care encounters for 140,000 patients from 2008 to 2013, including office visits, phone calls, and e-visits, all cholesterol tests, and all blood glucose tests for the physicians with the largest panel sizes. It was limited, however, to those patients who had three or more office visits over the period analyzed, as the study was designed to focus on active healthcare users.
The results were stark. After adopting e-visits - in this instance, essentially an email with a subject line and generic box of text - office visits increased by 6% as physicians met with patients who had reached out online. Physicians also ended up spending 45 more minutes each month on those visits. Oh, and the extra work of responding to patients requests did not bring extra compensation. "God knows what happens if you start paying doctors for these," Bavafa said. And with the increased workload came a corresponding 15% drop in the number of new patients physicians saw.
New county-level data from the CDC highlight the extreme geographic variation in opioid prescription rates, with some areas showing average morphine equivalents per capita 10 times greater than those of less-impacted counties. As seen in the map below, Appalachia appeared to be one giant hot spot. But every state had at least one county with a high per-capita rate of prescribed opioid use. Overall prescribing has fallen in the past few years, but the amount of opioids per person was still three times higher in 2015 than it was in 1999.
Those were the major findings in a Vital Signs release from the CDC, which focused on opioid prescribing. The data looked at trends through 2015, which notably does not include the agency's 2016 opioid guidelines. The numbers will serve as a baseline to measure the impact of those guidelines going forward, said acting CDC Director Anne Schuchat, MD. "The bottom line remains, we still have too many people getting these opioid prescriptions for too many days at too high a dose," she said.
In a statement, Patrice A. Harris, head of the American Medical Association's opioid task force, said she was pleased to see the report confirm that physicians have been making "more judicious" prescribing decisions. She also called for increased prescribing of naloxone and pointed to the AMA's own task force recommendations. Enough opioids were prescribed in 2015 to medicate every American for three weeks straight. To combat overprescribing, she said physicians should do so only when benefits outweigh risks. Behavioral and physical therapy and NSAIDs should be first choices. The average days' supply has increased annually for the past decade, from 13.3 in 2006 to 17.7 in 2015. If opioids are prescribed, the CDC recommends the duration of the prescription should be short - three days or less - and dose amounts should be as low as possible to achieve adequate pain relief.
In that way, the trendline is encouraging. The average daily morphine-equivalent dose per prescription has declined each year since 2006. Schuchat said she expects to see further reductions as the 2016 recommendations are implemented. "We won't be able to solve it overnight, but changes ... hold promise that prescribing practices can improve," she said.
Design issues and lack of alerts are two contributing factors
As electronic health records (EHR) usage grows more widespread, so too does the technology's role in malpractice claims, a new study found. The Doctors Company, a physician-owned medical malpractice insurer, found a continuous increase over the past decade in malpractice claims in which the use of EHRs contributed to patient injury. From 2007 through 2010, there were just two claims in which EHRs were a factor. From 2011 through December 2016, however, that number skyrocketed to 161, according to David B. Troxel, MD, medical director at The Doctors Company in Napa, California. The Doctors Company says this is its second study of EHR-related claims.Troxel noted in a statement that the EHR is typically a contributing factor in a claim, rather than the primary cause. he current study compares 66 claims made from July 2014 through December 2016 with the results of the first study of 97 claims from 2007 through June 2014.
Compared with the earlier research, the new study shows that system factors that contributed to claims increased 8%. These factors include technology and design issues, lack of integration of hospital EHR systems, and failure or lack of alerts and alarms. On the other hand, user factors, such as copy-and-paste errors, data entry errors, and alert fatigue, decreased 6%. Internal medicine, hospital medicine, and cardiology showed marked decreases among specialties involved in claims, while orthopedics, emergency medicine, and obstetrics/gynecology showed increases, the study found. The study also noted that hospital clinics/doctors' offices remain the top location for EHR-related claim events. Adoption of EHRs has been relatively fast. Data released last summer showed that only 4% of U.S. hospitals didn't use EHRs. The Doctors Company study stated that the technology "has great potential to advance both the practice of good medicine and patient safety." However, there are always unanticipated consequences when new technologies are rapidly adopted -- and the EHR is no exception," according to the study.
Vaccination against human papillomavirus (HPV) during pregnancy did not increase the risk of maternal or fetal complications, a retrospective, matched, case-control study showed. Comparison of 1,800 vaccine-exposed pregnancies and 7,000 unexposed pregnancies showed no significant differences in the frequency of six adverse outcomes. Unadjusted analyses did show higher rates of low birth weight, preterm birth, and major birth defects, but none of the differences remained significant in the propensity-matched analyses, reported Nikolai M. Scheller, MD, of the Statens Serum Institut in Copenhagen, and colleagues.
"Our results are consistent with other evidence that does not indicate that the vaccination of pregnant women with inactivated virus, bacterial, or toxoid vaccines generally confers a higher risk of adverse pregnancy outcomes than no such vaccination," they wrote in the New England Journal of Medicine. "Our results also confirm and considerably expand on results from previous studies of the quadrivalent HPV vaccine."
Five phase III, randomized clinical trials of the quadrivalent HPV vaccine, involving 3,500 women and 4,000 pregnancies, showed no increase in the risk of spontaneous abortion, stillbirth, or birth defects. However, most pregnancies occurred more than 6 months after exposure to the vaccine, precluding any direct assessment of the vaccine's risk during pregnancy, Scheller's group noted. Given the global recommendation for vaccination of girls and women, ages 9 to 26 years, against HPV, some women inevitably will have inadvertent exposure to the vaccine during early pregnancy. As the authors suggested, limited data have accumulated regarding the safety of the quadrivalent HPV vaccine during pregnancy.
In an effort to help fill the data void, investigators conducted a study that included identification of all pregnancies occurring in Denmark from Oct. 1, 2006 through Nov. 30, 2013. Using nationwide registry data, researchers collected information on vaccination, adverse pregnancy outcomes, and potential confounding factors among women included in data analysis. Scheller's group identified women who had exposure to the quadrivalent HPV vaccine (women who received the bivalent vaccine were excluded) during pre-specified time windows that varied according to the type of pregnancy complication: first trimester for birth defects; weeks 7 through 22 for spontaneous abortion; week 7 until birth for stillbirth; before 37 weeks of gestation for preterm birth; and any time during pregnancy for low birth weight and small for gestational age.
Subsequent data analysis included 463 vaccine-exposed women for spontaneous abortion; 1,665 for major birth defects; 501 for stillbirth; 1,774 for preterm birth; and 1,768 for low birth weight and small for gestational age. For propensity analyses, each vaccinated woman was matched with four unvaccinated women. The data showed that 65 birth defects in vaccine-exposed pregnancies versus 220 in unexposed pregnancies; 20 spontaneous abortions with vaccine exposure versus 131 without; 116 preterm births versus 407; 76 cases of low birth weight versus 277; 171 cases of small size for gestational age versus 783; and two stillbirths versus four. The absolute numbers translated into the following odds ratios for vaccine-exposed versus unexposed pregnancies:
Unadjusted analyses performed prior to propensity-score matching showed significantly increased risk associated with the quadrivalent HPV vaccine for low birth weight (OR 1.26, 95% CI 1.00-1.59), preterm birth (OR 1.38, 95% CI 1.14-1.67), and major birth defect (OR 1.36, 95% CI 1.06-1.75). Nonsignificant hazard ratios emerged from unadjusted analyses of spontaneous abortion (HR 0.93, 95% CI 0.60-1.44), small for gestational age (HR 0.98, 95% CI 0.83-1.14), and still birth (HR 1.90, 95% CI 0.48-7.61). The authors acknowledged several limitations of their study: potential for misclassification of physician-assigned pregnancy outcomes, possible unidentified confounding factors, and limited statistical power for some analyses because of small numbers of events.
Sleep disruption and perceived nighttime hot flashes trigger mild symptoms of depression during menopause, according to a new study published in the Journal of Clinical Endocrinology & Metabolism. Using a gonadotropin-releasing hormone agonist (GnRHa) experimental model to suppress estrogen levels in 29 premenopausal women, the researchers discovered that a worsening of mood was predicted by factors such as reductions in perceived sleep efficiency (all P<.045) and the number of nighttime (P<.006) hot flashes, but not the number of daytime (P<.28) hot flashes.
"Our results indicate that bothersome daytime hot flashes are not related to menopausal depressive symptoms -- which many people believed to be the case," Hadine Joffe, MD, MSc, Director of the Health at Brigham and Women's Hospital and Dana Farber Cancer Institute at Harvard Medical School said.
"The findings also show that it is not just the sleep problem related to the nighttime flashes that is linked with the mood problem. Rather, both nighttime hot flashes and sleep problems are of concern." The study included 29 healthy premenopausal women, ages 18 to 45, who reported no hot flashes, sleep disorders, or psychiatric illnesses. All women received a one-time dose of intramuscular leuprolide at 3.75 mg/day to rapidly induce hypoestrogenism and maintain ovarian suppression for the study period.
Depressive symptoms (using the Montgomery Åsberg Depression Rating Scale and the Beck Depression Inventory), sleep parameters (using two ambulatory polysomnography studies), subjective sleep quality, serum estradiol, and hot flashes were assessed both at baseline and after 4 weeks. Additionally, all women used a diary to record their hot flashes twice a day over the 4-week period - once in the morning to document the nighttime hot flashes and once in the evening to document the daytime flashes. Joffe and colleagues used an experimental model, rather than a naturalistic approach, in order to precisely assess the sequence of symptom onset.
"These common menopausal 'brain' symptoms of hot flashes, sleep disturbance, and depressive symptoms often present together, and it is hard to know which came first and which problem may be contributing to the other in naturalistic studies," she explained to MedPage Today. "This approach also enables us to subtract out each women's own baseline sleep and mood patterns in order to isolate the effect of hot flashes on sleep and mood." JoAnn E. Manson, MD, DrPH, chief of preventive medicine at Brigham and Women's Hospital and professor of medicine at Harvard Medical School, agreed that the method was successful for eliminating confounders and identifying a significant association. "Mood changes haven't been clearly linked to hot flashes or sleep quality in the past, and this study disentangles a lot of other potential contributing factors," she said.
In addition, "because the women are being compared with themselves, you don't have the potential confounding of aging or life circumstances that might contribute to symptoms in a natural menopause setting." Joffe too pointed out that "naturalistic studies are complicated by the variable amount of time that someone is symptomatic and the common co-occurrence of age-related changes in sleep that can be hard to separate out from menopause-related sleep changes." Nanette Santoro, MD, endowed chair in the Department of Obstetrics and Gynecology and professor in of Medicine in Aurora, called Joffe et al's method "an elegant approach, because it isolates the variable of estrogen and re-creates the problems that menopausal women often complain about."
And while she agreed that the study eliminated the confounding that often occurs around the time of natural menopause, she did note that "inducing 'medical menopause' with Lupron may not really mimic the same set of physiologic changes that happen in natural menopause, as the estrogen withdrawal is more rapid and perhaps more severe in this model." Joffe and colleagues concluded that women who reported having frequent nocturnal hot flashes were more likely to have mild symptoms of depression than those who reported fewer or no nocturnal hot flashes. Additionally, it was nighttime hot flashes, not those in the daytime, that were linked to depressive symptoms.
Manson said that one finding she found particularly interesting was that the subjective perception of nighttime hot flashes was more related to depressive symptoms than hot flashes that were objectively measured: "If women were not aware of being awakened by [the hot flashes], it may be less of a risk factor for depressive symptoms." One of the biggest takeaways from the results, the researchers said, was the importance of screening for mood disturbances during the menopausal transition and early menopause. "This [study] provides specific information to women and clinicians about the importance of treating nighttime hot flashes/night sweats as part of the approach to improving mood disturbance in this population," Joffe said. Manson agreed that the results suggest that clinicians should ask patients about nighttime hot flashes since they may be an important contributor to sleep disruption and to the development of depressive symptoms.
Joffe said that the small number of participants -- something often viewed as a limitation -- was actually one of the study's key advantages because "the variability is reduced since you can subtract each woman's own baseline profile and look at changes within each person." Another expert asked for her opinion, Diana L. Bitner, MD, NCMP, director of the Women's Health Network at Spectrum Health in Grand Rapids, Mich., said she agreed that the small study made sense: "[Joffe] was looking at each patient as their own control, done purposefully with the intent of understanding the evolution of symptoms as estrogen drops -- which allows for a smaller sample size."
Santoro also commented that the number of participants was well balanced against the extremely detailed nature of the study design. "The use of polysomnography and all the rating scales used, along with the multiple hormone assessments would be very, very difficult to do in a large sample. A larger sample with less detailed assessments might not necessarily be more convincing." Manson said that moving forward, further research is warranted, and ideally with a larger study. "One of the next types of studies to be done are what treatments are most effective in reducing hot flashes and improving sleep quality," she said. "It could look at the comparative effects of hormonal and nonhormonal therapies on nighttime hot flashes and poor sleep quality, as well as the depressive symptoms."