The heterogeneity of concentrated prescribing behavior: Theory and evidence from antipsychotics

Experimental evidence on the role of shared protocols as coordination device on clinical best practices

Our experiment assesses the level of coordination on clinical best practice among physicians and investigates whether the release of guidelines helps in supporting coordination. Based on three clinical vignettes using current national guidelines, physicians evaluate the appropriateness of each of the proposed courses of action. Afterwards, physicians are allowed to ask which action corresponds to national guidelines and change their ratings, if desired. On average, slightly more than half of the sample coordinated on appropriateness evaluations. Empirical analysis indicates that several organizational and individual variables influence the level of coordination. Additionally, the release of national guidelines improved both the level of conformity and coordination. Our findings suggest changes in implementation practices to increase the impact of these shared protocols in the health field.

Identifying depression and its determinants upon initiating treatment: ChatGPT versus primary care physicians

OBJECTIVE

To compare evaluations of depressive episodes and suggested treatment protocols generated by Chat Generative Pretrained Transformer (ChatGPT)-3 and ChatGPT-4 with the recommendations of primary care physicians.

METHODS

Vignettes were input to the ChatGPT interface. These vignettes focused primarily on hypothetical patients with symptoms of depression during initial consultations. The creators of these vignettes meticulously designed eight distinct versions in which they systematically varied patient attributes (sex, socioeconomic status (blue collar worker or white collar worker) and depression severity (mild or severe)). Each variant was subsequently introduced into ChatGPT-3.5 and ChatGPT-4. Each vignette was repeated 10 times to ensure consistency and reliability of the ChatGPT responses.

RESULTS

For mild depression, ChatGPT-3.5 and ChatGPT-4 recommended psychotherapy in 95.0% and 97.5% of cases, respectively. Primary care physicians, however, recommended psychotherapy in only 4.3% of cases. For severe cases, ChatGPT favoured an approach that combined psychotherapy, while primary care physicians recommended a combined approach. The pharmacological recommendations of ChatGPT-3.5 and ChatGPT-4 showed a preference for exclusive use of antidepressants (74% and 68%, respectively), in contrast with primary care physicians, who typically recommended a mix of antidepressants and anxiolytics/hypnotics (67.4%). Unlike primary care physicians, ChatGPT showed no gender or socioeconomic biases in its recommendations.

CONCLUSION

ChatGPT-3.5 and ChatGPT-4 aligned well with accepted guidelines for managing mild and severe depression, without showing the gender or socioeconomic biases observed among primary care physicians. Despite the suggested potential benefit of using atificial intelligence (AI) chatbots like ChatGPT to enhance clinical decision making, further research is needed to refine AI recommendations for severe cases and to consider potential risks and ethical issues.

Clinical Validity of a Machine Learning Decision Support System for Early Detection of Hepatitis B Virus: A Binational External Validation Study

HepB LiveTest is a machine learning decision support system developed for the early detection of hepatitis B virus (HBV). However, there is a lack of evidence on its generalisability. In this study, we aimed to externally assess the clinical validity and portability of HepB LiveTest in predicting HBV infection among independent patient cohorts from Nigeria and Australia. The performance of HepB LiveTest was evaluated by constructing receiver operating characteristic curves and estimating the area under the curve. Delong's method was used to estimate the 95% confidence interval (CI) of the area under the receiver-operating characteristic curve (AUROC). Compared to the Australian cohort, patients in the derivation cohort of HepB LiveTest and the hospital-based Nigerian cohort were younger (mean age, 45.5 years vs. 38.8 years vs. 40.8 years, respectively; p < 0.001) and had a higher incidence of HBV infection (1.9% vs. 69.4% vs. 57.3%). In the hospital-based Nigerian cohort, HepB LiveTest performed optimally with an AUROC of 0.94 (95% CI, 0.91-0.97). The model provided tailored predictions that ensured most cases of HBV infection did not go undetected. However, its discriminatory measure dropped to 0.60 (95% CI, 0.56-0.64) in the Australian cohort. These findings indicate that HepB LiveTest exhibits adequate cross-site transportability and clinical validity in the hospital-based Nigerian patient cohort but shows limited performance in the Australian cohort. Whilst HepB LiveTest holds promise for reducing HBV prevalence in underserved populations, caution is warranted when implementing the model in older populations, particularly in regions with low incidence of HBV infection.

There is no such thing as a validated prediction model

BACKGROUND

Clinical prediction models should be validated before implementation in clinical practice. But is favorable performance at internal validation or one external validation sufficient to claim that a prediction model works well in the intended clinical context?

MAIN BODY

We argue to the contrary because (1) patient populations vary, (2) measurement procedures vary, and (3) populations and measurements change over time. Hence, we have to expect heterogeneity in model performance between locations and settings, and across time. It follows that prediction models are never truly validated. This does not imply that validation is not important. Rather, the current focus on developing new models should shift to a focus on more extensive, well-conducted, and well-reported validation studies of promising models.

CONCLUSION

Principled validation strategies are needed to understand and quantify heterogeneity, monitor performance over time, and update prediction models when appropriate. Such strategies will help to ensure that prediction models stay up-to-date and safe to support clinical decision-making.

The formation of physician altruism

We study how patient-regarding altruism is formed by medical education. We elicit and structurally estimate altruistic preferences using experimental data from a large sample of medical students (N = 733) in Germany at different progress stages in their studies. The estimates reveal substantial heterogeneity in altruistic preferences of medical students. Patient-regarding altruism is highest for freshmen, significantly declines for students in the course of medical studies, and tends to increase again for last year students, who assist in clinical practice. Also, patient-regarding altruism is higher for females and positively associated to general altruism. Altruistic medical students have gained prior practical experience in healthcare, have lower income expectations, and are more likely to choose surgery and pediatrics as their preferred specialty.

Effects of utilization management on health outcomes: evidence from urinary tract infections and community-acquired pneumonia

BACKGROUND

Utilization management policies are pervasive in the Medicare Part D program. We assess the effect of utilization management restrictions in the Medicare Part D program on the quality of care in two clinical areas - community-acquired pneumonia (CAP) and urinary tract infections (UTI).

METHODS

In this study, we identified new cases of CAP and UTI from Medicare claims data from 2010 to 2016. We assessed the relationship between exposure to utilization management for antibiotic medications suitable for treating these conditions and adverse health outcomes, based on the Agency for Healthcare Research and Quality prevention quality indicators.

RESULTS

We identified 147,526 cases of CAP and 632,407 UTI cases in our data. In these samples, the adverse event rate varied from 3.6 to 5.7%. The probability of an adverse event increased by 0.75 (p = 0.061) percentage points for each ten percentage point increase in exposure to quantity limits (one form of utilization management) among people with CAP. There was no relationship between utilization management and adverse events in the UTI cohort.

CONCLUSIONS

In some circumstances, exposure to utilization management policies-particularly quantity limits-may adversely affect health.

The Effect of Biosimilar Prescription Targets for Erythropoiesis-Stimulating Agents on the Prescribing Behavior of Physicians in Germany

OBJECTIVES

This study aimed to aid decision makers by analyzing the impact of introducing biosimilar prescription targets on physician prescribing behavior in the prescription of biologic erythropoiesis-stimulating agents in Germany.

METHODS

We combined secondary data of regional level biosimilar prescription targets and secondary data of routinely collected claims data of dispensed prescriptions by physicians operating within the statutory health insurance system in ambulatory care across 7 German regions from 2009 to 2015. Two-way fixed-effects regression analysis was used to identify the average treatment effect of introducing biosimilar prescription targets at the physician level. The main outcome of interest was the share of biosimilar prescriptions on all prescriptions within the substance group. We compared 6 regions that introduced biosimilar prescription targets with 1 region without any prescription target policy.

RESULTS

Introducing biosimilar prescription targets increased the average share of biosimilars between 6 percentage points (P < .05) in Hamburg and up to 20 percentage points (P < .001) in Saxony-Anhalt. Stratification of specialists by prescription volume and adoption status indicated heterogeneous effects. We identified similar but higher effects for high-volume prescribers. Disentangling of effects with regard to the composition of biosimilar share suggested that the increase in biosimilar share was driven by increased biosimilar use accompanied by a nonsignificant decrease in original biologics prescriptions.

CONCLUSIONS

Prescription targets to alter physician prescribing behavior meet their intended goals by increasing biosimilar share. Physicians partly responded to the policy by decreasing overall prescriptions of the target substance. Prescription targets might be a useful tool, but decision makers need to consider all aspects of potential responses.

Identifying the effects of scientific information and recommendations on physicians' prescribing behavior

We investigate how the prescribing behavior of physicians reacts to scientific information and recommendations released by public authorities. Taking the example of antidepressant drugs, we use French panel data on exhaustive prescriptions made by a representative sample of general practitioners to more than 110,000 depressed patients between 2000 and 2008. New results revealing an increase in suicidal thinking among children taking selective serotonin reuptake inhibitors (SSRIs) were reported in 2004 and prompted the release of new guidelines by public health authorities. We identify the effect of this unexpected warning on physicians' drug choices while addressing the possibility that patients heterogeneity may be correlated with unobserved physician characteristics. While the warning decreased the average probability of prescribing SSRIs, we find that physicians' responses to the warning were very heterogeneous and larger if the physician had a higher preference for prescribing SSRIs before the warning.

Gender differences in physician decisions to adopt new prescription drugs

Physician adoption of new technologies is a key issue for population health and the sustainability of the healthcare system. This paper explores gender differences in general practitioners' (GPs) adoption of new oral anticoagulants. We combine detailed individual data on physician and practice style characteristics from the Medicine in Australia: Balancing Employment and Life (MABEL) panel survey of Australian physicians with administrative prescribing data from the Australian Pharmaceutical Benefits Scheme (PBS) and the Medicare Benefits Schedule (MBS) for the period January 1, 2012 and December 31, 2015. After adjusting for various factors proposed in the literature as drivers of this gender gap, in addition to risk preferences and personality traits, we find a large statistical gender difference in the speed of adoption with men being faster than women in uptake. However, conditional on having prescribed for the first time, female and male GPs differ only slightly in the intensity of use of these new drugs. We show that the gender gap depends on the measure of uptake and discuss possible channels that could be driving the relatively large gender difference that remains in the speed of adoption.

Generic entry of aromatase inhibitors and pharmaceutical access: Initiation of hormonal therapy, timeliness of initiation, and drug choice

BACKGROUND

The trade-offs between innovation and pharmaceutical access are central to the policy debate on drug pricing. High prices may limit access, result in medication underuse, and negatively affect outcomes. Generic drugs make treatments more affordable. Prior research measured access as utilization without a defined population that should receive certain drugs, it is unknown whether generic entry reduces underuse and thus improves access.

OBJECTIVES

To measure changes in access (use, timeliness) with the introduction of three generic aromatase inhibitors (AIs, oral breast cancer drugs) between June 2010 and June 2011.

METHODS

This population-based study included 93,650 older (65+) women diagnosed with hormone receptor-positive breast cancer between 2007 and 2013 in the Surveillance, Epidemiology and End Results-Medicare linked database. We examined changes in access with generic entry for initiation of any adjuvant hormonal therapy drug (AIs or tamoxifen) within one year of diagnosis, time from diagnosis to initiation, and choice of initial therapy.

RESULTS

Among 93,650 newly diagnosed breast cancer cases, 67,372 initiated one of the four drugs. With generic entry, initiation rates increased from 69.5% to 74.3%, but non-initiation remained high (up to 25.7%). After controlling for demographics, clinical factors, and insurance coverage, the probability of initiation increased by 4.6 percentage points (P < 0.001, 95%CI: [4.1,5.2]) after generic entry. With generic entry, estimated time to initiation decreased by 0.3 months (P < 0.001, 95%CI: [0.2,0.3]) from 4.1 months, and the probability of choosing AIs over tamoxifen increased by 5.9 percentage points (P < 0.001, 95%CI: [5.3,6.5]). Patterns did not substantially differ by level of cost-sharing.

CONCLUSIONS

Generic entry of AIs was associated with increased probability of receiving recommended treatments, timeliness of treatment, and the probability of receiving clinically preferred treatments. Price changes with generic entry only partially explained these improvements. High non-initiation rates after generic entry suggest prices are not the sole determinant of access.

The impact of drug quality ratings from health technology assessments on the adoption of new drugs by physicians in Germany

Payers are increasingly calling for the value of new drugs to be measured explicitly. We analyze how the availability of drug quality ratings by health technology assessment (HTA) agencies affects the adoption of new drugs by physicians in Germany. We combine data from drug quality ratings, promotional spending, and a physician panel. In a latent utility model, time to adoption is specified as a function of quality rating, promotional spending by manufacturers, and physician-specific variables. As expected, drugs with a positive rating were adopted faster (p < 0.001) than those without. However, our results suggest that it was the publication of the quality rating itself that affected adoption. Indeed, before a quality rating was published, drugs that went on to receive a positive quality rating were not adopted significantly faster than drugs that went on to receive a negative quality rating. In contrast, after the publication of the HTA quality rating, drugs with a positive rating were adopted significantly faster than those without (p < 0.05). The per physician value of a positive quality rating was EUR 393.50. Our results suggest that there are returns from HTAs beyond their use in price negotiations.

Treatment of mental illness in American adolescents varies widely within and across areas

Many mental health disorders first manifest in adolescence, and early treatment may affect the course of the disease. Using a large national database of insurance claims, this study focuses on variations in the type of care that adolescent patients receive when they are treated for an initial episode of mental illness. We found large variations in the probability that children receive follow-up care and in the type of follow-up care received across zip codes. We also found large variations in the probability that children receive drug treatments that raise a red flag when viewed through the lens of treatment guidelines: Overall, in the first 3 mo after their initial claim for mental illness, 44.85% of children who receive drug treatment receive benzodiazepines, tricyclic antidepressants, or a drug that is not Food and Drug Administration-approved for their age. On average, these children are 12 y old. While the supply of mental health professionals impacts treatment choices, little of the overall variation is explained by supply-side variables, and at least half of the variation in treatment outcomes occurs within zip codes. These results suggest that other factors, such as physician practice style, may play an important role in the types of treatment that children receive.

The definition and measurement of heterogeneity

Heterogeneity is an important concept in psychiatric research and science more broadly. It negatively impacts effect size estimates under case-control paradigms, and it exposes important flaws in our existing categorical nosology. Yet, our field has no precise definition of heterogeneity proper. We tend to quantify heterogeneity by measuring associated correlates such as entropy or variance: practices which are akin to accepting the radius of a sphere as a measure of its volume. Under a definition of heterogeneity as the degree to which a system deviates from perfect conformity, this paper argues that its proper measure roughly corresponds to the size of a system's event/sample space, and has units known as numbers equivalent. We arrive at this conclusion through focused review of more than 100 years of (re)discoveries of indices by ecologists, economists, statistical physicists, and others. In parallel, we review psychiatric approaches for quantifying heterogeneity, including but not limited to studies of symptom heterogeneity, microbiome biodiversity, cluster-counting, and time-series analyses. We argue that using numbers equivalent heterogeneity measures could improve the interpretability and synthesis of psychiatric research on heterogeneity. However, significant limitations must be overcome for these measures-largely developed for economic and ecological research-to be useful in modern translational psychiatric science.

Physician-Level Cost Control Measures and Regional Variation of Biosimilar Utilization in Germany

Biologic drugs represent a large and growing portion of health expenditures. Increasing the use of biosimilars is a promising option for controlling spending growth in pharmaceutical care. Amid the considerable uncertainty concerning physicians’ decision to prescribe biosimilars, explicit cost control measures may help increase biosimilar use. We analyze the role of regional cost control measures for biosimilars and their association with physician prescriptions in ambulatory care in Germany. We collect data on cost control measures implemented by German physician associations and national claims data on statutory health insurance covering 2009 to 2015. We perform panel regressions that include time and physician fixed effects to identify the average associations between cost control measures and biosimilar share/use while controlling for unobserved physician heterogeneity, patient structure, and socioeconomic factors. We identify 44 measures (priority prescribing, biosimilar quota) for erythropoiesis-stimulating substances, filgrastim, and somatropin. Estimates of cost control measures and their consequences for biosimilar share and use are heterogeneous by drug, measure type, and physician group. Across specialists, biosimilar quotas accounted for 5.13% to 9.75% of the total average biosimilar share of erythropoiesis-stimulating substances. Explicit quota regulations are more effective than priority prescribing. Regional variation in biosimilar use can be partly attributed to the presence of cost control measures.

UNDERSTANDING DOCTOR DECISION MAKING: THE CASE OF DEPRESSION TREATMENT

Treatment for depression is complex, requiring decisions that may involve trade-offs between exploiting treatments with the highest expected value and experimenting with treatments with higher possible payoffs. Using patient claims data, we show that among skilled doctors, using a broader portfolio of drugs predicts better patient outcomes, except in cases where doctors' decisions violate loose professional guidelines. We introduce a behavioral model of decision making guided by our empirical observations. The model's novel feature is that the trade-off between exploitation and experimentation depends on the doctor's diagnostic skill. The model predicts that higher diagnostic skill leads to greater diversity in drug choice and better matching of drugs to patients even among doctors with the same initial beliefs regarding drug effectiveness. Consistent with the finding that guideline violations predict poorer patient outcomes, simulations of the model suggest that increasing the number of possible drug choices can lower performance.

Functional Neuroimaging in Psychiatry-Aiding in Diagnosis and Guiding Treatment. What the American Psychiatric Association Does Not Know

While early efforts in psychiatry were focused on uncovering the neurobiological basis of psychiatric symptoms, they made little progress due to limited ability to observe the living brain. Today, we know a great deal about the workings of the brain; yet, none of this neurobiological awareness has translated into the practice of psychiatry. The categorical system which dominates psychiatric diagnosis and thinking fails to match up to the real world of genetics, sophisticated psychological testing, and neuroimaging. Nevertheless, the American Psychiatric Association (APA) recently published a position paper stating that neuroimaging provided no benefit to the diagnosis and treatment of psychiatric disorders. Using the diagnosis of depression as a model, we illustrate how setting aside the unrealistic expectation of a pathognomonic "fingerprint" for categorical diagnoses, we can avoid missing the biological and, therefore, treatable contributors to psychopathology which can and are visualized using functional neuroimaging. Infection, toxicity, inflammation, gut-brain dysregulation, and traumatic brain injury can all induce psychiatric manifestations which masquerade as depression and other psychiatric disorders. We review these and provide illustrative clinical examples. We further describe situations for which single photon emission computed tomography (SPECT) and positron emission tomography (PET) functional neuroimaging already meet or exceed the criteria set forth by the APA to define a neuroimaging biomarker, including the differential diagnosis of Alzheimer's disease and other dementias, the differential diagnosis of ADHD, and the evaluation of traumatic brain injury. The limitations, both real and perceived, of SPECT and PET functional neuroimaging in the field of psychiatry are also elaborated. An important overarching concept for diagnostic imaging in all its forms, including functional neuroimaging, is that imaging allows a clinician to eliminate possibilities, narrow the differential diagnosis, and tailor the treatment plan. This progression is central to any medical diagnostic process.

Physician Beliefs and Patient Preferences: A New Look at Regional Variation in Health Care Spendingf

There is considerable controversy about the causes of regional variations in health care expenditures. Using vignettes from patient and physician surveys linked to fee-for-service Medicare expenditures, this study asks whether patient demand-side factors or physician supply-side factors explain these variations. The results indicate that patient demand is relatively unimportant in explaining variations. Physician organizational factors matter, but the most important factor is physician beliefs about treatment. In Medicare, we estimate that 35 percent of spending for end-of-life care and 12 percent of spending for heart attack patients (and for all enrollees) is associated with physician beliefs unsupported by clinical evidence. (JEL D83, H75, I11, I18).