Is an online skin cancer toolkit an effective way to educate primary care physicians about skin cancer diagnosis and referral?

BACKGROUND: Skin disorders account for over 20% of GP consultations. Half of dermatology referrals to secondary care are for skin lesions, but only 12% of urgent skin cancer referrals are deemed appropriate. Suitably designed online learning resources may positively impact GP confidence in the recognition of skin cancer and improve patient outcomes. OBJECTIVE: This study evaluated the impact of a national, online, skin cancer recognition toolkit on GP confidence and knowledge in diagnosing skin cancers and referral behaviour to secondary care. METHODS: The toolkit, consisting of a referral decision aid, lesion recognition resource, clinical cases and a quiz, was launched in March 2012. Website usage statistics and online focus groups were used to assess the usability of the website and perceived changes in behaviour. The impact of the toolkit was assessed using national skin cancer referral data, cross-sectional questionnaires and urgent skin cancer referral data to two NHS trusts. RESULTS: The toolkit was accessed by 20% of GPs in England from 20th March to 31st October 2012; spending a mean of over 5 minutes each, with over 33% return users. A survey revealed that the toolkit improved perceptions of skin cancer training and self-reported knowledge about skin cancer referral pathways. Analysis of referral patterns did not identify an impact of the toolkit on number or appropriateness of urgent skin cancer referrals in the eight months following the launch of the website. Online focus groups confirmed the usefulness of the resource and suggested a positive influence on knowledge and referral behaviour. CONCLUSION: The skin cancer toolkit is an accessible online learning resource for improving confidence with skin cancer referral amongst GPs. Although we were unable to identify any immediate changes in skin cancer diagnoses or appropriate referral behaviours, research is required to evaluate its longer term effects on outcomes.

Differences in clinical decision making between internists and cardiologists.

BACKGROUND: Whether cardiologists or internists use discretionary tests differently for noncritical cardiological presentation is unclear. OBJECTIVE: To explore differences in decision making for 3 common scenarios. METHODS: We asked 318 cardiologists and 598 internists to manage scenario patients presenting with (1) uncomplicated syncope, (2) nonanginal chest pain, and (3) nonspecific electrocardiographic changes. Participants also estimated baseline clinical risk for each scenario and answered questions on uncertainty, malpractice concerns, and cost consciousness. We used chi 2 analysis, analysis of variance, and t tests to compare management choice and test ordering. Response rate was 50%. RESULTS: Initial management choices (ie, admit or discharge, allow or delay surgery) were similar but subsequent testing differed substantially. For a 50-year-old woman with uncomplicated syncope, cardiologists more often recommended cardiological tests such as exercise treadmill tests (37% vs 18%, 95% confidence interval [CI] for difference: 10%-28%) and signal-averaged electrocardiograms (13% vs 4%, 95% CI for difference: 3%-15%) but less often requested neurological tests (29% vs 37%, 95% CI for difference: -17% to 1%). For a 42-year-old man with nonanginal chest pain, cardiologists more frequently ordered exercise tests (70% vs 51%, 95% CI for difference: 10%-28%). For a 53-year-old woman with nonspecific electrocardiographic changes, equal proportions of cardiologists and internists ordered exercise tests (56%) but cardiologists recommended thallium studies more often (73% vs 47%, 95% CI for difference: 10%-36%). For all scenarios, average charges for diagnostic evaluations by cardiologists and internists were similar. CONCLUSIONS: In 3 noncritical cardiology scenarios, discretionary test use by cardiologists and internists differed substantially, although this was not reflected in dollar resources. Internists tended toward a broader diagnostic evaluation while cardiologists tended to focus on cardiological tests. The potential effect on clinical outcomes is unknown.

Physicians' personal malpractice experiences are not related to defensive clinical practices.

Whether personal malpractice experience is part of a tort signal prompting physicians to practice defensively is unclear. To explore this issue further, we assessed how physicians' malpractice experiences affect clinical decision making. We surveyed 1,540 physicians from four specialty groups (cardiologists, surgeons, obstetrician-gynecologists, and internists) using specialty-specific clinical scenarios. Physicians were in active private practice, were covered by a single malpractice insurer for five or more years, and worked in an eastern state. The net response rate was 54 percent (835 of 1,540) but measurable bias, based on practice characteristics, was negligible. Physicians evaluated clinical scenarios that were designed to maximize potential for finding positive defensive practices (extra tests and procedures). Then they rated how various factors influenced their decisions and answered questions on practice attitudes. The study compared management and testing recommendations among physicians with varying levels of malpractice exposure, which we defined in three separate ways. Participants were unaware of the study hypotheses. Physicians with greater malpractice experience showed no systematic differences in initial management choice or subsequent test recommendations. For example, similar percentages of internists in the top and bottom claims rate quartiles admitted a patient with syncope (78 percent versus 73 percent; p = 42), discharged a patient with nonspecific chest pain (80 percent versus 80 percent; p = .88), and delayed surgery in a patient with nonspecific changes on a electrocardiograph (58 percent versus 68 percent; p = .18). Attitudes about malpractice also did not differ with varying malpractice experience. Personal malpractice experience is not a predominant factor in the tort signal that prompts physicians to engage in defensive practices, to the extent that such practices exist.

Empirical Bayes methods for estimating hospital-specific mortality rates.

We present alternative methods for estimating hospital-level mortality rates to those used by the Health Care Finance Administration for Medicare patients. We use an empirical Bayes model to represent the different sources of variation in observed hospital-specific mortality rates and we use a logistic regression model to adjust for severity differences (in patient mix) across hospitals. In addition to providing a principled derivation of a standard error for the commonly used estimator, our fully model-based formulation produces much more accurate estimates and resolves the severe problem of multiple comparisons that arises when extreme estimates are used to identify exceptional hospitals. We estimate models for each of four disease conditions using the national Medicare mortality data base which does not contain patient severity descriptors, and mortality data from national samples which do include patient severity descriptors. We find substantial between-hospital variation in the unadjusted death rates from the national data base. Mortality rates differ substantially with patient severity in our models, but the sample sizes are too small to yield reliable estimates of the between-hospital variation in adjusted mortality rates.

Amending the National Practitioner Data Bank reporting requirements: are small claims predictive of large claims?

This study addresses whether a physician incurring small malpractice claims is predictive of large claims. This is one consideration behind reevaluating whether all claims that result in an indemnity payment should continue to be reported to the National Practitioner Data Bank, or whether claims with payments below some "floor" should be excluded. Using a claims database from 3,098 physicians for 1977-1986, both cross-sectional and longitudinal analyses show that an individual having a small claim (under $30,000) is indicative of a propensity to incur large claims. This finding is robust to the cutpoint between large and small claims.

Malpractice claims data as a quality improvement tool. I. Epidemiology of error in four specialties.

OBJECTIVE: --To identify potentially preventable sources of medical injury in obstetrics and gynecology, general surgery, anesthesiology, and radiology. DESIGN: --Retrospective review of physician malpractice claim records. SETTING: --Large New Jersey physician malpractice insurer. PARTICIPANTS: --Physicians practicing obstetrics and gynecology, general surgery, anesthesiology, and radiology and covered by the insurance carrier during any portion of 1977 through 1989. MAIN OUTCOME MEASURES: --Proportion of claims due to negligence associated with errors in (1) patient management, (2) technical performance, and (3) medical and nursing staff coordination and the clinical and financial consequences of such errors. RESULTS: --Among 1371 claims ascribed to negligence, patient management errors were cited most frequently in all four specialties (48% to 75%) and, compared with performance and coordination problems, were generally associated with a higher frequency of serious injury and higher median payments. Coordination problems accounted for about 9% of claims. In obstetrics and gynecology, newborn delivery claims usually arose from management errors (57% to 68%), whereas gynecologic procedure claims were most often associated with performance errors (55% to 73%). Underperformance of cesarean section was cited more frequently than overperformance (31% vs 3%). General surgery claims were about equally divided between management and performance types regardless of procedure. Failure to perform appropriate diagnostic testing or monitoring was the main problem in 3% to 8% of claims. CONCLUSION: --Malpractice data can be used to identify problem-prone clinical processes and suggest interventions that may reduce negligence.

Malpractice claims data as a quality improvement tool. II. Is targeting effective?

OBJECTIVE: --To evaluate the usefulness of malpractice claims data for identifying (1) physicians who are prone to negligent errors and (2) physician and hospital characteristics associated with particular kinds of errors. DESIGN: --Retrospective review of physician malpractice claim records. SETTING: --Large New Jersey physician malpractice insurer. PARTICIPANTS: --Physicians practicing obstetrics and gynecology, general surgery, anesthesiology, or radiology and covered by the insurance carrier for any portion of 1977 through 1989. MAIN OUTCOME MEASURES: --Claims were classified into 11 clinical error categories comprising three broad groups: patient management problems, technical performance problems, and staff coordination problems. Outcomes were expressed as per-physician frequency of claims due to negligence and proportion of claims associated with various types of errors. RESULTS: --Using 5 years of claims history to predict long-term claims proneness was more accurate than chance alone by 57% in obstetrics and gynecology, 33% in general surgery, 11% in anesthesiology, and 15% in radiology. Cross-validated recursive partitioning showed that among physician characteristics, only specialty was predictive of physician error profiles. For physician claims arising in acute care hospitals, hospital size and location in addition to hospital services discriminated among different error profiles; the cross-validated accuracy of this method was 69% compared with 22% accuracy achieved by random prediction. CONCLUSION: --Use of physicians' malpractice claims histories to target individuals for education or sanctions is problematic because of the only modest predictive power of such claims histories.

The demand for episodes of treatment in the Health Insurance Experiment.

This paper analyzes claims data from the RAND Insurance Experiment, which were grouped into episodes of treatment. The insurance plans in the experiment have coinsurance and a cap on out-of-pocket spending. Using new statistical techniques to adjust for the increased sickliness of those who exceed the cap, the effects of coinsurance on cost per episode and number of episodes are estimated. Cost sharing reduced the number of episodes but had little effect on cost per episode. People in the experiment responded myopically as their current insurance status changed through the year. The price elasticity of spending was about -0.2 throughout the range of coinsurance studied. When data permit it, the study of episodes complements analyses of annual medical spending by revealing more about how decisions to spend are made within the year.

Use, costs, and quality of medical services: impact of the New Mexico peer review system. A 1971-1975 study.

To assess the effect of areawide peer review (such as that conducted by Professional Standards Review Organizations [PSROs]) on use, cost and quality of medical services, we evaluated 4 years of data on the efforts of the New Mexico Experimental Medical Care Review Organization in reviewing medical services for the Medicaid population. Utilization review had no demonstrable impact on hospital use; hospital days per 100 eligible persons rose 5.0% and 43.4% for persons enrolled all 4 years in Aid to Families with Dependent Children (AFDC) and Aid to the Permanently and Totally Disabled, respectively. Peer review produced no net dollar savings; over 4 years, the amount paid for all services per AFDC-eligible person rose 85%. Peer review improved the quality of ambulatory care through large reductions (75%) in medically unnecessary injections. If these findings are replicated elsewhere, they suggest that the goal of the PSRO program to control costs by curtailing utilization may be difficult to achieve, the quality of care goal may be pursued successfully, and the PSRO mission should be focused more on the latter.