Including Information on Overdiagnosis in Shared Decision Making: A Review of Prostate Cancer Screening Decision Aids

Background. Overdiagnosis is an accepted harm of cancer screening, but studies of prostate cancer screening decision aids have not examined provision of information important in communicating the risk of overdiagnosis, including overdiagnosis frequency, competing mortality risk, and the high prevalence of indolent cancers in the population. Methods. We undertook a comprehensive review of all publicly available decision aids for prostate cancer screening, published in (or translated to) the English language, without date restrictions. We included all decision aids from a recent systematic review and screened excluded studies to identify further relevant decision aids. We used a Google search to identify further decision aids not published in peer reviewed medical literature. Two reviewers independently screened the decision aids and extracted information on communication of overdiagnosis. Disagreements were resolved through discussion or by consulting a third author. Results. Forty-one decision aids were included out of the 80 records identified through the search. Most decision aids (n = 32, 79%) did not use the term overdiagnosis but included a description of it (n = 38, 92%). Few (n = 7, 17%) reported the frequency of overdiagnosis. Little more than half presented the benefits of prostate cancer screening before the harms (n = 22, 54%) and only 16, (39%) presented information on competing risks of mortality. Only 2 (n = 2, 5%) reported the prevalence of undiagnosed prostate cancer in the general population. Conclusion. Most patient decision aids for prostate cancer screening lacked important information on overdiagnosis. Specific guidance is needed on how to communicate the risks of overdiagnosis in decision aids, including appropriate content, terminology and graphical display.

Highlights

Most patient decision aids for prostate cancer screening lacks important information on overdiagnosis.Specific guidance is needed on how to communicate the risks of overdiagnosis.

Learning from the pandemic: mortality trends and seasonality of deaths in Australia in 2020

AIM

To assess whether the observed numbers and seasonality of deaths in Australia during 2020 differed from expected trends based on 2015-19 data.

METHODS

We used provisional death data from the Australian Bureau of Statistics, stratified by state, age, sex and cause of death. We compared 2020 deaths with 2015-19 deaths using interrupted time series adjusted for time trend and seasonality. We measured the following outcomes along with 95% confidence intervals: observed/expected deaths (rate ratio: RR), change in seasonal variation in mortality (amplitude ratio: AR) and change in week of peak seasonal mortality (phase difference: PD).

RESULTS

Overall 4% fewer deaths from all causes were registered in Australia than expected in 2020 [RR 0·96 (0·95-0·98)] with reductions across states, ages and sex strata. There were fewer deaths from respiratory illness [RR 0·79 (0·76-0·83)] and dementia [RR 0·95 (0·93-0·98)] but more from diabetes [RR 1·08 (1·04-1·13)]. Seasonal variation was reduced for deaths overall [AR 0·94 (0·92-0·95)], and for deaths due to respiratory illnesses [AR 0·78 (0·74-0·83)], dementia [AR 0.92 (0.89-0.95)] and ischaemic heart disease [0.95 (0.90-0.97)].

CONCLUSIONS

The observed reductions in respiratory and dementia deaths and the reduced seasonality in ischaemic heart disease deaths may reflect reductions in circulating respiratory (non-SARS-CoV-2) pathogens resulting from the public health measures taken in 2020. The observed increase in diabetes deaths is unexplained and merits further study.

Can patient-led surveillance detect subsequent new primary or recurrent melanomas and reduce the need for routinely scheduled follow up? Statistical analysis plan for the MEL-SELF randomised controlled trial

BACKGROUND

The MEL-SELF trial is a randomised controlled trial of patient-led surveillance compared to clinician-led surveillance in people treated for localised cutaneous melanoma (stage 0, I, II). The primary trial aim is to determine if patient led-surveillance compared to clinician-led surveillance increases the proportion of participants who are diagnosed with a new primary or recurrent melanoma at a fast-tracked unscheduled clinic visit. The secondary outcomes include time to diagnosis of any skin cancer, psychosocial outcomes, acceptability, and resource use.

OBJECTIVE

The objective of this report is to outline and publish the pre-determined statistical analysis plan before the database lock and the start of analysis.

METHODS/DESIGN

The statistical analysis plan describes the overall analysis principles, including how participants will be included in each analysis, the presentation of the results, adjustments for covariates, the primary and secondary outcomes, and their respective analyses. In addition, we present the planned sensitivity and subgroup analyses. A separate analysis plan will be published for health economic outcomes.

RESULTS

The MEL-SELF statistical analysis plan has been designed to minimize bias in estimating effects of the intervention on primary and secondary outcomes. By pre-specifying analyses, we ensure the study's integrity and believability while enabling the reproducibility of the final analysis.

CONCLUSION

This detailed statistical analysis plan will help to ensure transparency of reporting of results from the MEL-SELF trial.

TRIAL REGISTRATION

Australian New Zealand Clinical Trials Registry (ANZCTR): ACTRN12621000176864. Registered 18 February 2021, https://www.anzctr.org.au/ACTRN12621000176864.aspx.

Evaluation of the Incremental Value of a Coronary Artery Calcium Score Beyond Traditional Cardiovascular Risk Assessment: A Systematic Review and Meta-analysis

IMPORTANCE

Coronary artery calcium scores (CACS) are used to help assess patients' cardiovascular status and risk. However, their best use in risk assessment beyond traditional cardiovascular factors in primary prevention is uncertain.

OBJECTIVE

To find, assess, and synthesize all cohort studies that assessed the incremental gain from the addition of a CACS to a standard cardiovascular disease (CVD) risk calculator (or CVD risk factors for a standard calculator), that is, comparing CVD risk score plus CACS with CVD risk score alone.

EVIDENCE REVIEW

Eligible studies needed to be cohort studies in primary prevention populations that used 1 of the CVD risk calculators recommended by national guidelines (Framingham Risk Score, QRISK, pooled cohort equation, NZ PREDICT, NORRISK, or SCORE) and assessed and reported incremental discrimination with CACS for estimating the risk of a future cardiovascular event.

FINDINGS

From 2772 records screened, 6 eligible cohort studies were identified (with 1043 CVD events in 17 961 unique participants) from the US (n = 3), the Netherlands (n = 1), Germany (n = 1), and South Korea (n = 1). Studies varied in size from 470 to 5185 participants (range of mean [SD] ages, 50 [10] to 75.1 [7.3] years; 38.4%-59.4% were women). The C statistic for the CVD risk models without CACS ranged from 0.693 (95% CI, 0.661-0.726) to 0.80. The pooled gain in C statistic from adding CACS was 0.036 (95% CI, 0.020-0.052). Among participants classified as being at low risk by the risk score and reclassified as at intermediate or high risk by CACS, 85.5% (65 of 76) to 96.4% (349 of 362) did not have a CVD event during follow-up (range, 5.1-10.0 years). Among participants classified as being at high risk by the risk score and reclassified as being at low risk by CACS, 91.4% (202 of 221) to 99.2% (502 of 506) did not have a CVD event during follow-up.

CONCLUSIONS AND RELEVANCE

This systematic review and meta-analysis found that the CACS appears to add some further discrimination to the traditional CVD risk assessment equations used in these studies, which appears to be relatively consistent across studies. However, the modest gain may often be outweighed by costs, rates of incidental findings, and radiation risks. Although the CACS may have a role for refining risk assessment in selected patients, which patients would benefit remains unclear. At present, no evidence suggests that adding CACS to traditional risk scores provides clinical benefit.

What is the prevalence of fear of cancer recurrence in cancer survivors and patients? A systematic review and individual participant data meta-analysis

OBJECTIVE

Care for fear of cancer recurrence (FCR) is considered the most common unmet need among cancer survivors. Yet the prevalence of FCR and predisposing factors remain inconclusive. To support targeted care, we provide a comprehensive overview of the prevalence and severity of FCR among cancer survivors and patients, as measured using the short form of the validated Fear of Cancer Recurrence Inventory (FCRI-SF). We also report on associations between FCR and clinical and demographic characteristics.

METHODS

This is a systematic review and individual participant data (IPD) meta-analysis on the prevalence of FCR. In the review, we included all studies that used the FCRI-SF with adult (≥18 years) cancer survivors and patients. Date of search: 7 February 2020. Risk of bias was assessed using the Joanna Briggs Institute critical appraisal tool.

RESULTS

IPD were requested from 87 unique studies and provided for 46 studies comprising 11,226 participants from 13 countries. 9311 respondents were included for the main analyses. On the FCRI-SF (range 0-36), 58.8% of respondents scored ≥13, 45.1% scored ≥16 and 19.2% scored ≥22. FCR decreased with age and women reported more FCR than men. FCR was found across cancer types and continents and for all time periods since cancer diagnosis.

CONCLUSIONS

FCR affects a considerable number of cancer survivors and patients. It is therefore important that healthcare providers discuss this issue with their patients and provide treatment when needed. Further research is needed to investigate how best to prevent and treat FCR and to identify other factors associated with FCR. The protocol was prospectively registered (PROSPERO CRD42020142185).

Assessing the Potential for Patient-led Surveillance After Treatment of Localized Melanoma (MEL-SELF): A Pilot Randomized Clinical Trial

IMPORTANCE

Patient-led surveillance is a promising new model of follow-up care following excision of localized melanoma.

OBJECTIVE

To determine whether patient-led surveillance in patients with prior localized primary cutaneous melanoma is as safe, feasible, and acceptable as clinician-led surveillance.

DESIGN, SETTING, AND PARTICIPANTS

This was a pilot for a randomized clinical trial at 2 specialist-led clinics in metropolitan Sydney, Australia, and a primary care skin cancer clinic managed by general practitioners in metropolitan Newcastle, Australia. The participants were 100 patients who had been treated for localized melanoma, owned a smartphone, had a partner to assist with skin self-examination (SSE), and had been routinely attending scheduled follow-up visits. The study was conducted from November 1, 2018, to January 17, 2020, with analysis performed from September 1, 2020, to November 15, 2020.

INTERVENTION

Participants were randomized (1:1) to 6 months of patient-led surveillance (the intervention comprised usual care plus reminders to perform SSE, patient-performed dermoscopy, teledermatologist assessment, and fast-tracked unscheduled clinic visits) or clinician-led surveillance (the control was usual care).

MAIN OUTCOMES AND MEASURES

The primary outcome was the proportion of eligible and contacted patients who were randomized. Secondary outcomes included patient-reported outcomes (eg, SSE knowledge, attitudes, and practices, psychological outcomes, other health care use) and clinical outcomes (eg, clinic visits, skin surgeries, subsequent new primary or recurrent melanoma).

RESULTS

Of 326 patients who were eligible and contacted, 100 (31%) patients (mean [SD] age, 58.7 [12.0] years; 53 [53%] men) were randomized to patient-led (n = 49) or clinician-led (n = 51) surveillance. Data were available on patient-reported outcomes for 66 participants and on clinical outcomes for 100 participants. Compared with clinician-led surveillance, patient-led surveillance was associated with increased SSE frequency (odds ratio [OR], 3.5; 95% CI, 0.9 to 14.0) and thoroughness (OR, 2.2; 95% CI, 0.8 to 5.7), had no detectable adverse effect on psychological outcomes (fear of cancer recurrence subscale score; mean difference, -1.3; 95% CI, -3.1 to 0.5), and increased clinic visits (risk ratio [RR], 1.5; 95% CI, 1.1 to 2.1), skin lesion excisions (RR, 1.1; 95% CI, 0.6 to 2.0), and subsequent melanoma diagnoses and subsequent melanoma diagnoses (risk difference, 10%; 95% CI, -2% to 23%). New primary melanomas and 1 local recurrence were diagnosed in 8 (16%) of the participants in the intervention group, including 5 (10%) ahead of routinely scheduled visits; and in 3 (6%) of the participants in the control group, with none (0%) ahead of routinely scheduled visits (risk difference, 10%; 95% CI, 2% to 19%).

CONCLUSIONS AND RELEVANCE

This pilot of a randomized clinical trial found that patient-led surveillance after treatment of localized melanoma appears to be safe, feasible, and acceptable. Experiences from this pilot study have prompted improvements to the trial processes for the larger trial of the same intervention.

TRIAL REGISTRATION

http://anzctr.org.au Identifier: ACTRN12616001716459.

Rethinking Low-Risk Papillary Thyroid Cancers < 1cm (Papillary Microcarcinomas): An Evidence Review for Recalibrating Diagnostic Thresholds and/or Alternative Labels

Background: Recalibrating diagnostic thresholds or using alternative labels may mitigate overdiagnosis and overtreatment of papillary microcarcinoma (mPTC). We aimed at identifying and collating relevant epidemiological evidence on mPTC, to assess the case for recalibration and/or new labels. Methods: We searched EMBASE and PubMed databases from inception to December 2020 for natural history, autopsy, diagnostic drift, and diagnostic reproducibility studies. Where a relevant systematic review was pre-identified, only new articles were additionally included. Non-English articles were excluded. One author screened titles and abstracts. Two authors screened full text articles, performed quality assessments, and extracted data. We undertook narrative synthesis of included evidence (pooled estimates from systematic reviews and single estimates from primary studies). Results: One systematic review of patients undergoing active surveillance found that after 5 years of follow-up, 5.3% (95% confidence interval [CI 4.4-6.4%]) of the mPTC lesions had increased in size by ≥3 mm, and 1.6% [CI 1.1-2.4%] of patients had lymph node metastases. Among 7 new primary studies (including 3 updates on 2 studies included in the systematic review), 1-5% of patients undergoing active surveillance had lymph node metastases after a median follow-up of 1-10 years. One systematic review found that subclinical thyroid cancer incidentally discovered at autopsy is relatively common, with a pooled prevalence of 11.2% [CI 6.7-16.1%] among studies that examined the whole thyroid. Four diagnostic drift studies evaluated the new classification of non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP). Three studies of cases previously diagnosed as papillary thyroid cancer found 1.3-2.3% were reclassified as NIFTP (reclassifications were from follicular variation of papillary thyroid cancer [FVPTC]). One study of 48 cases previously diagnosed as mPTC found that 23.5% were reclassified as NIFTP. Thirteen reproducibility studies of papillary thyroid lesions found substantial variation in the histopathological diagnosis of thyroid lesions, including FVPTC and NIFTP classifications (no study evaluated mPTC). Conclusions: This review supports consideration of recalibrating diagnostic thresholds and/or alternative labels for low-risk mPTC.

Setting analytical performance specifications using HbA1c as a model measurand

Analytical performance specifications (APS) for measurands describe the minimum analytical quality requirements for their measurement. These APS are used to monitor and contain the systematic (trueness/bias) and random errors (precision/imprecision) of a laboratory measurement to ensure the results are "fit for purpose" in informing clinical decisions about managing a patient's health condition. In this review, we highlighted the wide variation in the setting of APS, using different levels of evidence, as recommended by the Milan Consensus, and approaches. The setting of a priori defined outcome-based APS for HbA1c remains challenging. Promising indirect alternatives seek to link the clinical utility of HbA1c and APS by defining statistical confidence for interpreting the laboratory values, or through simulation of clinical performance at varying levels of analytical performance. APS defined based on biological variation estimates in healthy individuals using the current formulae are unachievable by nearly all routine laboratory methods for HbA1c testing. On the other hand, the APS employed in external quality assurance programs have been progressively tightened, and greatly facilitate the improved quality of HbA1c testing. Laboratories should select the APS that fits their intended clinical use and should document the data and rationale underpinning those selections. Where possible common APS should be adopted across a region or country to facilitate the movement of patients and patient data across health care facilities.

Evaluation of Gender Inequity in Thyroid Cancer Diagnosis: Differences by Sex in US Thyroid Cancer Incidence Compared With a Meta-analysis of Subclinical Thyroid Cancer Rates at Autopsy

IMPORTANCE

Thyroid cancer is more common in women than in men, but the associated causes of these differences are not fully understood.

OBJECTIVE

To compare sex-specific thyroid cancer rates in the US to the prevalence of subclinical thyroid cancer at autopsy.

DATA SOURCES

Data on thyroid cancer incidence and mortality by sex among US adults (≥18 years) were extracted from the National Cancer Institute's Surveillance, Epidemiology, and End Results Program (SEER) data for 1975 to 2017. Embase, PubMed, and Web of Science databases were searched for studies on the prevalence of subclinical thyroid cancer at autopsy of men and women, from inception to May 31, 2021.

STUDY SELECTION

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline was used to perform a systematic search for articles reporting the prevalence of subclinical thyroid cancer in autopsy results of both women and men. Of 101 studies identified, 8 studies containing 12 data sets met inclusion criteria; ie, they examined the whole thyroid gland, stated the number of thyroids examined, and reported results by sex. Excluded studies reported thyroid cancer in Japan after the atomic bombs or Chernobyl after the nuclear disaster; did not examine the whole thyroid gland or had incomplete information on thyroid examination methods; or did not report rates by sex.

DATA EXTRACTION AND SYNTHESIS

Thyroid cancer incidence and mortality data by sex, histologic type, and tumor size were extracted from SEER. The inverse variance heterogeneity model was used to meta-analyze the prevalence and the odds ratio of subclinical thyroid cancer by sex from 8 studies (12 data sets) on thyroid cancer prevalence in autopsy results.

MAIN OUTCOMES AND MEASURES

Incidence and mortality of thyroid cancer, by histologic type and tumor size; prevalence of thyroid cancer in autopsy results.

RESULTS

In 2017, 90% of thyroid cancers diagnosed were papillary thyroid cancer (PTC) and in 2013 to 2017, the women to men incidence ratio for small (≤2 cm) PTC was 4.39:1. The incidence ratio approached 1:1 as cancer type lethality increased. The ratio of thyroid cancer mortality by gender was 1.02:1 and remained stable from 1992 to 2017. Results of the meta-analysis showed that the pooled autopsy prevalence of subclinical PTC was 14% in women (95% CI, 8%-20%) and 11% in men (95% CI, 5%-18%). The pooled odds ratio of subclinical PTC in women compared with men was 1.07 (95% CI, 0.80-1.42).

CONCLUSIONS AND RELEVANCE

This cohort study and meta-analysis found that the belief that women get thyroid cancer more often than men is an oversimplification. The gender disparity is mostly confined to the detection of small subclinical PTCs, which are equally common in both sexes at autopsy but identified during life much more often in women than men. As the lethality of the cancer type increases, the ratio of detection by gender approaches 1:1. This phenomenon may be associated with gender differences in health care utilization and patterns of clinical thinking and can harm both women, who are subject to overdetection, and men, who may be at risk of underdetection.

Can patient-led surveillance detect subsequent new primary or recurrent melanomas and reduce the need for routinely scheduled follow-up? A protocol for the MEL-SELF randomised controlled trial

BACKGROUND

Most subsequent new primary or recurrent melanomas might be self-detected if patients are trained to systematically self-examine their skin and have access to timely medical review (patient-led surveillance). Routinely scheduled clinic visits (clinician-led surveillance) is resource-intensive and has not been shown to improve health outcomes; fewer visits may be possible if patient-led surveillance is shown to be safe and effective. The MEL-SELF trial is a randomised controlled trial comparing patient-led surveillance with clinician-led surveillance in people who have been previously treated for localised melanoma.

METHODS

Stage 0/I/II melanoma patients (n = 600) from dermatology, surgical, or general practice clinics in NSW Australia, will be randomised (1:1) to the intervention (patient-led surveillance, n = 300) or control (usual care, n = 300). Patients in the intervention will undergo a second randomisation 1:1 to polarised (n = 150) or non-polarised (n = 150) dermatoscope. Patient-led surveillance comprises an educational booklet, skin self-examination (SSE) instructional videos; 3-monthly email/SMS reminders to perform SSE; patient-performed dermoscopy with teledermatologist feedback; clinical review of positive teledermoscopy through fast-tracked unscheduled clinic visits; and routinely scheduled clinic visits following each clinician's usual practice. Clinician-led surveillance comprises an educational booklet and routinely scheduled clinic visits following each clinician's usual practice. The primary outcome, measured at 12 months, is the proportion of participants diagnosed with a subsequent new primary or recurrent melanoma at an unscheduled clinic visit. Secondary outcomes include time from randomisation to diagnosis (of a subsequent new primary or recurrent melanoma and of a new keratinocyte cancer), clinicopathological characteristics of subsequent new primary or recurrent melanomas (including AJCC stage), psychological outcomes, and healthcare use. A nested qualitative study will include interviews with patients and clinicians, and a costing study we will compare costs from a societal perspective. We will compare the technical performance of two different models of dermatoscope (polarised vs non-polarised).

DISCUSSION

The findings from this study may inform guidance on evidence-based follow-up care, that maximises early detection of subsequent new primary or recurrent melanoma and patient wellbeing, while minimising costs to patients, health systems, and society.

TRIAL REGISTRATION

Australian New Zealand Clinical Trials Registry (ANZCTR): ACTRN12621000176864 . Registered on 18 February 2021.

Estimating the potential impact of interventions to reduce over-calling and under-calling of melanoma

BACKGROUND

Pathologists sometimes disagree over the histopathologic diagnosis of melanoma. 'Over-calling' and 'under-calling' of melanoma may harm individuals and healthcare systems.

OBJECTIVES

To estimate the extent of 'over-calling' and 'under-calling' of melanoma for a population undergoing one excision per person and to model the impact of potential solutions.

METHODS

In this epidemiological modelling study, we undertook simulations using published data on the prevalence and diagnostic accuracy of melanocytic histopathology in the U.S.

POPULATION

We simulated results for 10 000 patients each undergoing excision of one melanocytic lesion, interpreted by one community pathologist. We repeated the simulation using a hypothetical intervention that improves diagnostic agreement between community pathologist and a specialist dermatopathologist. We then evaluated four scenarios for how melanocytic lesions judged to be neither clearly benign (post-test probability of melanoma < 5%), nor clearly malignant (post-test probability of melanoma > 90%) might be handled, before sending for expert dermatopathologist review to decide the final diagnosis. These were (1) no intervention before expert review, (2) formal second community pathologist review, (3) intervention to increase diagnostic agreement and (4) both the intervention and formal second community pathologist review. The main outcomes were the probability of 'over-calling' and 'under-calling' melanoma, and number of lesions requiring expert referral for each scenario.

RESULTS

For 10 000 individuals undergoing excision of one melanocytic lesion, interpreted by a community pathologist, a hypothetical intervention to improve histopathology agreement reduced the number of benign lesions 'over-called' as melanoma from 308 to 164 and the number of melanomas 'under-called' from 289 to 240. If all uncertain diagnoses were sent for expert review, the number of referrals would decrease from 1500 to 737 cases if formal second community pathologist review was used, and to 701 cases if the hypothetical intervention was additionally used.

CONCLUSIONS

Interventions to improve histopathology agreement may reduce melanoma 'over-calling' and 'under-calling'.

Mendelian Randomization in Cardiovascular Research: Establishing Causality When There Are Unmeasured Confounders

Mendelian randomization is an epidemiological approach to making causal inferences using observational data. It makes use of the natural randomization that occurs in the generation of an individual's genetic makeup in a way that is analogous to the study design of a randomized controlled trial and uses instrumental variable analysis where the genetic variant(s) are the instrument (analogous to random allocation to treatment group in an randomized controlled trial). As with any instrumental variable, there are 3 assumptions that must be made about the genetic instrument: (1) it is associated (not necessarily causally) with the exposure (relevance condition); (2) it is associated with the outcome only through the exposure (exclusion restriction condition); and (3) it does not share a common cause with the outcome (ie, no confounders of the genetic instrument and outcome, independence condition). Using the example of type II diabetes and coronary artery disease, we demonstrate how the method may be used to investigate causality and discuss potential benefits and pitfalls. We conclude that although Mendelian randomization studies can usually not establish causality on their own, they may usefully contribute to the evidence base and increase our certainty about the effectiveness (or otherwise) of interventions to reduce cardiovascular disease.

A methods review of posttrial follow-up studies of cardiovascular prevention finds potential biases in estimating legacy effects

OBJECTIVES

The objective of the study was to assess the methods used, and potential for bias, in posttrial studies of cardiovascular disease (CVD) where legacy effects may be estimated.

STUDY DESIGN AND SETTING

We undertook a methods review of posttrial studies after randomized controlled trials (RCTs) of interventions to prevent CVD. For each included article, we extracted information on important aspects of the design and analysis of the study, and on the reporting of legacy effects.

RESULTS

Of 2,622 retrieved articles, 46 were included in the review: 13 on blood glucose control, 13 on blood pressure control, and 20 on blood lipid control. The median duration for the RCT and posttrial follow-up studies was 5.0 and 5.7 years, respectively. At least 80% of initial RCT participants were enrolled in the posttrial study in 32 of the reports. Most reports used both linkage to routine administrative data sources and active data collection for the posttrial study. Of the 46 included articles, the authors assessed and reported posttrial covariate balance in 29 and made statistical adjustments in the analysis for potential confounding in 25. Posttrial results were reported separately to overall results (from time of randomization) in 21 articles. Legacy effects were claimed in 19 reports, of which 16 could be justified on the basis of the posttrial results.

CONCLUSION

Posttrial studies may provide valuable information for investigating legacy effects, but better reporting of results is needed to realize their full potential. Robust methods of data collection and analysis may address the risk of selection and confounding biases in posttrial studies.

Why clinicians overtest: development of a thematic framework

BACKGROUND

Medical tests provide important information to guide clinical management. Overtesting, however, may cause harm to patients and the healthcare system, including through misdiagnosis, false positives, false negatives and overdiagnosis. Clinicians are ultimately responsible for test requests, and are therefore ideally positioned to prevent overtesting and its unintended consequences. Through this narrative literature review and workshop discussion with experts at the Preventing Overdiagnosis Conference (Sydney, 2019), we aimed to identify and establish a thematic framework of factors that influence clinicians to request non-recommended and unnecessary tests.

METHODS

Articles exploring factors affecting clinician test ordering behaviour were identified through a systematic search of MedLine in April 2019, forward and backward citation searches and content experts. Two authors screened abstract titles and abstracts, and two authors screened full text for inclusion. Identified factors were categorised into a preliminary framework which was subsequently presented at the PODC for iterative development.

RESULTS

The MedLine search yielded 542 articles; 55 were included. Another 10 articles identified by forward-backward citation and content experts were included, resulting in 65 articles in total. Following small group discussion with workshop participants, a revised thematic framework of factors was developed: "Intrapersonal" - fear of malpractice and litigation; clinician knowledge and understanding; intolerance of uncertainty and risk aversion; cognitive biases and experiences; sense of medical obligation "Interpersonal" - pressure from patients and doctor-patient relationship; pressure from colleagues and medical culture; "Environment/context" - guidelines, protocols and policies; financial incentives and ownership of tests; time constraints, physical vulnerabilities and language barriers; availability and ease of access to tests; pre-emptive testing to facilitate subsequent care; contemporary medical practice and new technology CONCLUSION: This thematic framework may raise awareness of overtesting and prompt clinicians to change their test request behaviour. The development of a scale to assess clinician knowledge, attitudes and practices is planned to allow evaluation of clinician-targeted interventions to reduce overtesting.

Legacy effect of fibrate add-on therapy in diabetic patients with dyslipidemia: a secondary analysis of the ACCORDION study

BACKGROUND

The Action to Control Cardiovascular Risk in Diabetes (ACCORD)-Lipid study found no evidence of a beneficial effect of statin-fibrate combined treatment, compared to statins alone, on cardiovascular outcomes and mortality in type 2 diabetes mellitus after 5 years of active treatment. However, a beneficial reduction in major CVD events was shown in a pre-specified sub-group of participants with dyslipidemia. The extended follow-up of this trial provides the opportunity to further investigate possible beneficial effects of fibrates in this group of patients. We aimed to evaluate possible "legacy effects" of fibrate add-on therapy on mortality and major cardiovascular outcomes in patients with dyslipidemia.

METHODS

The ACCORD-lipid study was a randomized controlled trial of 5518 participants assigned to receive simvastatin plus fenofibrate vs simvastatin plus placebo. After randomized treatment allocation had finished at the end of the trial, all surviving participants were invited to attend an extended follow-up study (ACCORDION) to continue prospective collection of clinical outcomes. We undertook a secondary analysis of trial and post-trial data in patients who had dyslipidemia. The primary outcome was all-cause and cardiovascular mortality, and secondary outcomes were nonfatal myocardial infarction, stroke, congestive heart failure and major coronary heart disease. We used an intention-to-treat approach to analysis to make comparisons between the original randomized treatment groups.

RESULTS

853 participants with dyslipidemia had survived at the end of the trial. Most participants continued to use statins, but few used fibrates in either group during the post-trial period. The incidence rates in the fenofibrate group were lower with respect to all-cause mortality, CVD mortality, nonfatal myocardial infarction, congestive heart failure and major coronary heart disease than those in the placebo group over a post-trial follow-up. Allocation to the combined fibrate-statin treatment arm during the trial period had a beneficial legacy effect on all-cause mortality (adjusted HR = 0.65, 95% CI 0.45-0.94; P = 0.02).

CONCLUSIONS

Fibrate treatment during the initial trial period was associated with a legacy benefit of improved survival over a post-trial follow-up. These findings support re-evaluation of fibrates as an add-on strategy to statins in order to reduce cardiovascular risk in diabetic patients with dyslipidemia. Trial registration clinicaltrials.gov, Identifier: NCT00000620.

Correction: Prognostic impact of systolic blood pressure variability in people with diabetes

[This corrects the article DOI: 10.1371/journal.pone.0194084.].