Assessing Performance and Clinical Usefulness in Prediction Models With Survival Outcomes: Practical Guidance for Cox Proportional Hazards Models.

Risk prediction models need thorough validation to assess their performance. Validation of models for survival outcomes poses challenges due to the censoring of observations and the varying time horizon at which predictions can be made. This article describes measures to evaluate predictions and the potential improvement in decision making from survival models based on Cox proportional hazards regression.As a motivating case study, the authors consider the prediction of the composite outcome of recurrence or death (the "event") in patients with breast cancer after surgery. They developed a simple Cox regression model with 3 predictors, as in the Nottingham Prognostic Index, in 2982 women (1275 events over 5 years of follow-up) and externally validated this model in 686 women (285 events over 5 years). Improvement in performance was assessed after the addition of progesterone receptor as a prognostic biomarker.The model predictions can be evaluated across the full range of observed follow-up times or for the event occurring by the end of a fixed time horizon of interest. The authors first discuss recommended statistical measures that evaluate model performance in terms of discrimination, calibration, or overall performance. Further, they evaluate the potential clinical utility of the model to support clinical decision making according to a net benefit measure. They provide SAS and R code to illustrate internal and external validation.The authors recommend the proposed set of performance measures for transparent reporting of the validity of predictions from survival models.

Early pregnancy outcomes of IVF cycles using donor versus partner sperm: analysis of 1 376 454 cycles recorded by the Human Fertilisation and Embryology Authority (1991-2016).

STUDY QUESTION: Are the early pregnancy outcomes of IVF pregnancies conceived with donor sperm different to those conceived with partner sperm? SUMMARY ANSWER: Pregnancies conceived with donor sperm have a lower odds of early pregnancy loss and ectopic pregnancy compared to pregnancies conceived with partner sperm. WHAT IS KNOWN ALREADY: The number of cycles using donor sperm has risen significantly in recent years. Adverse early pregnancy outcomes have a negative impact on women and their partners. The evidence available to date regarding early pregnancy outcomes for pregnancies conceived with IVF donor sperm is limited by low numbers and lower-quality studies. STUDY DESIGN, SIZE, DURATION: This is a retrospective cohort study of 1 376 454 cycles conceived with either donor or partner sperm between 1991 and 2016 as recorded in the Human Fertilisation and Embryology Authority (HFEA) Register. PARTICIPANTS/MATERIALS, SETTING, METHODS: The HFEA has recorded data on all fertility treatments carried out in the UK from 1991 onwards, and it publishes this data in an anonymized form. This study assessed the outcomes of all pregnancies conceived with donor sperm and compared them to those conceived with partner sperm among IVF cycles recorded in the HFEA anonymized dataset from 1991 to 2016. Cycles that included intrauterine insemination, donor oocytes, preimplantation genetic testing, oocyte thaw cycles and alternative fertility treatments were excluded. The outcomes of interest were biochemical pregnancy, miscarriage, ectopic pregnancy, stillbirth and live birth. Logistic regression was used to adjust for confounding factors including age of the female partner, cause of infertility, history of previous pregnancy, fresh or frozen cycle, IVF or ICSI, number of embryos transferred, and year of treatment. Results are reported as adjusted odds ratios (aOR) and 95% CIs. MAIN RESULTS AND THE ROLE OF CHANCE: This study found reductions in the odds of biochemical pregnancy (aOR 0.82, 95% CI 0.78-0.86), miscarriage (aOR 0.93, 95% CI 0.89-0.97), and ectopic pregnancy (aOR 0.77, 95% CI 0.66-0.90) among pregnancies as a result of the use of donor sperm as opposed to partner sperm. LIMITATIONS, REASONS FOR CAUTION: This study is retrospective and limited by the constraints of routinely collected data. No data were available for maternal characteristics such as BMI, smoking and partner age, which could all be potential confounders. Clustering of multiple pregnancies within women could not be accounted for as the data are reported only at the cycle level with no maternal identifiers. WIDER IMPLICATIONS OF THE FINDINGS: This study has demonstrated that there are no increased risks of adverse pregnancy outcome with donor sperm pregnancies. The reduction in miscarriage in pregnancies using donor sperm suggests that sperm could have a role in miscarriage, as the selection process for being accepted as donor is stringent. STUDY FUNDING/COMPETING INTEREST(S): No external funding was sought for this study. C.A. has received funding from Ferring to attend a UK meeting for trainees in reproductive Medicine. A.M. has received funding from Ferring, Cook, Merck Serono, Geodon Ritcher, and Pharmasure for speaking at, or attending, meetings relating to reproductive medicine. She has also participated in a Ferring advisory board. S.B. has received grants from Tenovus and the UK Medical Research Council. She has also been supported with a Medical Research Scotland PhD studentship. TRIAL REGISTRATION NUMBER: N/A.

External validation of models for predicting cumulative live birth over multiple complete cycles of IVF treatment.

STUDY QUESTION: Can two prediction models developed using data from 1999 to 2009 accurately predict the cumulative probability of live birth per woman over multiple complete cycles of IVF in an updated UK cohort? SUMMARY ANSWER: After being updated, the models were able to estimate individualized chances of cumulative live birth over multiple complete cycles of IVF with greater accuracy. WHAT IS KNOWN ALREADY: The McLernon models were the first to predict cumulative live birth over multiple complete cycles of IVF. They were converted into an online calculator called OPIS (Outcome Prediction In Subfertility) which has 3000 users per month on average. A previous study externally validated the McLernon models using a Dutch prospective cohort containing data from 2011 to 2014. With changes in IVF practice over time, it is important that the McLernon models are externally validated on a more recent cohort of patients to ensure that predictions remain accurate. STUDY DESIGN, SIZE, DURATION: A population-based cohort of 91 035 women undergoing IVF in the UK between January 2010 and December 2016 was used for external validation. Data on frozen embryo transfers associated with these complete IVF cycles conducted from 1 January 2017 to 31 December 2017 were also collected. PARTICIPANTS/MATERIALS, SETTING, METHODS: Data on IVF treatments were obtained from the Human Fertilisation and Embryology Authority (HFEA). The predictive performances of the McLernon models were evaluated in terms of discrimination and calibration. Discrimination was assessed using the c-statistic and calibration was assessed using calibration-in-the-large, calibration slope, and calibration plots. Where any model demonstrated poor calibration in the validation cohort, the models were updated using intercept recalibration, logistic recalibration, or model revision to improve model performance. MAIN RESULTS AND THE ROLE OF CHANCE: Following exclusions, 91 035 women who underwent 144 734 complete cycles were included. The validation cohort had a similar distribution age profile to women in the development cohort. Live birth rates over all complete cycles of IVF per woman were higher in the validation cohort. After calibration assessment, both models required updating. The coefficients of the pre-treatment model were revised, and the updated model showed reasonable discrimination (c-statistic: 0.67, 95% CI: 0.66 to 0.68). After logistic recalibration, the post-treatment model showed good discrimination (c-statistic: 0.75, 95% CI: 0.74 to 0.76). As an example, in the updated pre-treatment model, a 32-year-old woman with 2 years of primary infertility has a 42% chance of having a live birth in the first complete ICSI cycle and a 77% chance over three complete cycles. In a couple with 2 years of primary male factor infertility where a 30-year-old woman has 15 oocytes collected in the first cycle, a single fresh blastocyst embryo transferred in the first cycle and spare embryos cryopreserved, the estimated chance of live birth provided by the post-treatment model is 46% in the first complete ICSI cycle and 81% over three complete cycles. LIMITATIONS, REASONS FOR CAUTION: Two predictors from the original models, duration of infertility and previous pregnancy, which were not available in the recent HFEA dataset, were imputed using data from the older cohort used to develop the models. The HFEA dataset does not contain some other potentially important predictors, e.g. BMI, ethnicity, race, smoking and alcohol intake in women, as well as measures of ovarian reserve such as antral follicle count. WIDER IMPLICATIONS OF THE FINDINGS: Both updated models show improved predictive ability and provide estimates which are more reflective of current practice and patient case mix. The updated OPIS tool can be used by clinicians to help shape couples' expectations by informing them of their individualized chances of live birth over a sequence of multiple complete cycles of IVF. STUDY FUNDING/COMPETING INTEREST(S): This study was supported by an Elphinstone scholarship scheme at the University of Aberdeen and Aberdeen Fertility Centre, University of Aberdeen. S.B. has a commitment of research funding from Merck. D.J.M. and M.B.R. declare support for the present manuscript from Elphinstone scholarship scheme at the University of Aberdeen and Assisted Reproduction Unit at Aberdeen Fertility Centre, University of Aberdeen. D.J.M. declares grants received by University of Aberdeen from NHS Grampian, The Meikle Foundation, and Chief Scientist Office in the past 3 years. D.J.M. declares receiving an honorarium for lectures from Merck. D.J.M. is Associate Editor of Human Reproduction Open and Statistical Advisor for Reproductive BioMed Online. S.B. declares royalties from Cambridge University Press for a book. S.B. declares receiving an honorarium for lectures from Merck, Organon, Ferring, Obstetric and Gynaecological Society of Singapore, and Taiwanese Society for Reproductive Medicine. S.B. has received support from Merck, ESHRE, and Ferring for attending meetings as speaker and is on the METAFOR and CAPRE Trials Data Monitoring Committee. TRIAL REGISTRATION NUMBER: N/A.

Predicting cumulative live birth for couples beginning their second complete cycle of in vitro fertilization treatment.

STUDY QUESTION: Can we develop an IVF prediction model to estimate individualized chances of a live birth over multiple complete cycles of IVF in couples embarking on their second complete cycle of treatment? SUMMARY ANSWER: Yes, our prediction model can estimate individualized chances of cumulative live birth over three additional complete cycles of IVF. WHAT IS KNOWN ALREADY: After the completion of a first complete cycle of IVF, couples who are unsuccessful may choose to undergo further treatment to have their first child, while those who have had a live birth may decide to have more children. Existing prediction models can estimate the overall chances of success in couples before commencing IVF but are unable to revise these chances on the basis of the couple's response to a first treatment cycle in terms of the number of eggs retrieved and pregnancy outcome. This makes it difficult for couples to plan and prepare emotionally and financially for the next step in their treatment. STUDY DESIGN, SIZE, DURATION: For model development, a population-based cohort was used of 49 314 women who started their second cycle of IVF including ICSI in the UK from 1999 to 2008 using their own oocytes and their partners' sperm. External validation was performed on data from 39 442 women who underwent their second cycle from 2010 to 2016. PARTICIPANTS/MATERIALS, SETTING, METHODS: Data about all UK IVF treatments were obtained from the Human Fertilisation and Embryology Authority (HFEA) database. Using a discrete time logistic regression model, we predicted the cumulative probability of live birth from the second up to and including the fourth complete cycles of IVF. Inverse probability weighting was used to account for treatment discontinuation. Discrimination was assessed using c-statistic and calibration was assessed using calibration-in-the-large and calibration slope. MAIN RESULTS AND THE ROLE OF CHANCE: Following exclusions, 49 314 women with 73 053 complete cycles were included. 12 408 (25.2%) had a live birth resulting from their second complete cycle. Cumulatively, 17 394 (35.3%) had a live birth over complete cycles two to four. The model showed moderate discriminative ability (c-statistic: 0.65, 95% CI: 0.64 to 0.65) and evidence of overprediction (calibration-in-the-large = -0.08) and overfitting (calibration slope 0.85, 95% CI: 0.81 to 0.88) in the validation cohort. However, after recalibration the fit was much improved. The recalibrated model identified the following key predictors of live birth: female age (38 versus 32 years-adjusted odds ratio: 0.59, 95% CI: 0.57 to 0.62), number of eggs retrieved in the first complete cycle (12 versus 4 eggs; 1.34, 1.30 to 1.37) and outcome of the first complete cycle (live birth versus no pregnancy; 1.78, 1.66 to 1.91; live birth versus pregnancy loss; 1.29, 1.23 to 1.36). As an example, a 32-year-old with 2 years of non-tubal infertility who had 12 eggs retrieved from her first stimulation and had a live birth during her first complete cycle has a 46% chance of having a further live birth from the second complete cycle of IVF and an 81% chance over a further three cycles. LIMITATIONS, REASONS FOR CAUTION: The developed model was updated using validation data that was 6 to 12 years old. IVF practice continues to evolve over time, which may affect the accuracy of predictions from the model. We were unable to adjust for some potentially important predictors, e.g. BMI, smoking and alcohol intake in women, as well as measures of ovarian reserve such as antral follicle count. These were not available in the linked HFEA dataset. WIDER IMPLICATIONS OF THE FINDINGS: By appropriately adjusting for couples who discontinue treatment, our novel prediction model will provide more realistic chances of live birth in couples starting a second complete cycle of IVF. Clinicians can use these predictions to inform discussion with couples who wish to plan ahead. This prediction tool will enable couples to prepare emotionally, financially and logistically for IVF treatment. STUDY FUNDING/COMPETING INTEREST(S): This study was supported by an Elphinstone scholarship scheme at the University of Aberdeen and Aberdeen Fertility Centre, University of Aberdeen. The authors have no conflict of interest. TRIAL REGISTRATION NUMBER: N/A.

Validation of Risk Prediction Models to Inform Clinical Decisions After Acute Kidney Injury.

RATIONALE & OBJECTIVE: There is limited evidence to guide follow-up after acute kidney injury (AKI). Knowledge gaps include which patients to prioritize, at what time point, and for mitigation of which outcomes. In this study, we sought to compare the net benefit of risk model-based clinical decisions following AKI. STUDY DESIGN: External validation of 2 risk models of AKI outcomes: the Grampian -Aberdeen (United Kingdom) AKI readmissions model and the Alberta (Canada) kidney disease risk model of chronic kidney disease (CKD) glomerular (G) filtration rate categories 4 and 5 (CKD G4 and G5). Process mining to delineate existing care pathways. SETTING & PARTICIPANTS: Validation was based on data from adult hospital survivors of AKI from Grampian, 2011-2013. PREDICTORS: KDIGO-based measures of AKI severity and comorbidities specified in the original models. OUTCOMES: Death or readmission within 90 days for all hospital survivors. Progression to new CKD G4-G5 for patients surviving at least 90 days after AKI. ANALYTICAL APPROACH: Decision curve analysis to assess the "net benefit" of use of risk models to guide clinical care compared to alternative approaches (eg, prioritizing all AKI, severe AKI, or only those without kidney recovery). RESULTS: 26,575 of 105,461 hospital survivors in Grampian (mean age, 60.9 ± 19.8 [SD] years) were included for validation of the death or readmission model, and 9,382 patients (mean age, 60.9 ± 19.8 years) for the CKD G4-G5 model. Both models discriminated well (area under the curve [AUC], 0.77 and 0.86, respectively). Decision curve analysis showed greater net benefit for follow up of all AKI than only severe AKI in most cases. Both original and refitted models provided net benefit superior to any other decision strategy. In process mining of all hospital discharges, 41% of readmissions and deaths occurred among people recovering after AKI. 1,464 of 3,776 people (39%) readmitted after AKI had received no intervening monitoring. LIMITATIONS: Both original models overstated risks, indicating a need for regular updating. CONCLUSIONS: Follow up after AKI has potential net benefit for preempting readmissions, death, and subsequent CKD progression. Decisions could be improved by using risk models and by focusing on AKI across a full spectrum of severity. The current lack of monitoring among many with poor outcomes indicates possible opportunities for implementation of decision support.

Prioritizing IVF treatment in the post-COVID 19 era: a predictive modelling study based on UK national data.

STUDY QUESTION: Can we use prediction modelling to estimate the impact of coronavirus disease 2019 (COVID 19) related delay in starting IVF or ICSI in different groups of women? SUMMARY ANSWER: Yes, using a combination of three different models we can predict the impact of delaying access to treatment by 6 and 12 months on the probability of conception leading to live birth in women of different age groups with different categories of infertility. WHAT IS KNOWN ALREADY: Increased age and duration of infertility can prejudice the chances of success following IVF, but couples with unexplained infertility have a chance of conceiving naturally without treatment whilst waiting for IVF. The worldwide suspension of IVF could lead to worse outcomes in couples awaiting treatment, but it is unclear to what extent this could affect individual couples based on age and cause of infertility. STUDY DESIGN, SIZE, DURATION: A population-based cohort study based on national data from all licensed clinics in the UK obtained from the Human Fertilisation and Embryology Authority Register. Linked data from 9589 women who underwent their first IVF or ICSI treatment in 2017 and consented to the use of their data for research were used to predict livebirth. PARTICIPANTS/MATERIALS, SETTING, METHODS: Three prediction models were used to estimate the chances of livebirth associated with immediate treatment versus a delay of 6 and 12 months in couples about to embark on IVF or ICSI. MAIN RESULTS AND THE ROLE OF CHANCE: We estimated that a 6-month delay would reduce IVF livebirths by 0.4%, 2.4%, 5.6%, 9.5% and 11.8% in women aged <30, 30-35, 36-37, 38-39 and 40-42 years, respectively, while corresponding values associated with a delay of 12 months were 0.9%, 4.9%, 11.9%, 18.8% and 22.4%, respectively. In women with known causes of infertility, worst case (best case) predicted chances of livebirth after a delay of 6 months followed by one complete IVF cycle in women aged <30, 30-35, 36-37, 38-39 and 40-42 years varied between 31.6% (35.0%), 29.0% (31.6%), 23.1% (25.2%), 17.2% (19.4%) and 10.3% (12.3%) for tubal infertility and 34.3% (39.2%), 31.6% (35.3%) 25.2% (28.5%) 18.3% (21.3%) and 11.3% (14.1%) for male factor infertility. The corresponding values in those treated immediately were 31.7%, 29.8%, 24.5%, 19.0% and 11.7% for tubal factor and 34.4%, 32.4%, 26.7%, 20.2% and 12.8% in male factor infertility. In women with unexplained infertility the predicted chances of livebirth after a delay of 6 months followed by one complete IVF cycle were 41.0%, 36.6%, 29.4%, 22.4% and 15.1% in women aged <30, 30-35, 36-37, 38-39 and 40-42 years, respectively, compared to 34.9%, 32.5%, 26.9%, 20.7% and 13.2% in similar groups of women treated without any delay. The additional waiting period, which provided more time for spontaneous conception, was predicted to increase the relative number of babies born by 17.5%, 12.6%, 9.1%, 8.4% and 13.8%, in women aged <30, 30-35, 36-37, 38-39 and 40-42 years, respectively. A 12-month delay showed a similar pattern in all subgroups. LIMITATIONS, REASONS FOR CAUTION: Major sources of uncertainty include the use of prediction models generated in different populations and the need for a number of assumptions. Although the models are validated and the bases for the assumptions are robust, it is impossible to eliminate the possibility of imprecision in our predictions. Therefore, our predicted live birth rates need to be validated in prospective studies to confirm their accuracy. WIDER IMPLICATIONS OF THE FINDINGS: A delay in starting IVF reduces success rates in all couples. For the first time, we have shown that while this results in fewer babies in older women and those with a known cause of infertility, it has a less detrimental effect on couples with unexplained infertility, some of whom conceive naturally whilst waiting for treatment. Post-COVID 19, clinics planning a phased return to normal clinical services should prioritize older women and those with a known cause of infertility. STUDY FUNDING/COMPETING INTEREST(S): No external funding was received for this study. B.W.M. is supported by an NHMRC Practitioner Fellowship (GNT1082548) and reports consultancy work for ObsEva, Merck, Merck KGaA, Guerbet and iGenomics. S.B. is Editor-in-Chief of Human Reproduction Open. None of the other authors declare any conflicts of interest. TRIAL REGISTRATION NUMBER: N/A.

Prevalence of PErioperAtive CHildhood obesitY in children undergoing general anaesthesia in the UK: a prospective, multicentre, observational cohort study.

BACKGROUND: Childhood obesity has become a serious global healthcare challenge. No UK data currently define its anaesthetic and perioperative implications. We aimed to determine obesity prevalence amongst UK children undergoing general anaesthesia and the incidence of predefined adverse perioperative events, and to compare perioperative obesity rates with National Child Measurement Programme (NCMP) data. METHODS: During a site-selected consecutive 7-day study period, all children (2-16 yr) undergoing general anaesthesia were included. Anonymised hospital, surgical, and procedural details; demographic data; and adverse perioperative events were collected prospectively by Paediatric Anaesthesia Trainee Research Network (PATRN) collaborators. RESULTS: For this study, 102 UK hospitals participated and 4232 cases were included in the final analysis; 76% of hospitals did not routinely calculate BMI. In addition, 3030 (71.6%; 95% confidence interval [CI]: 70.2-73.0%) children of healthy weight were compared with 537 (12.7%; 11.7-13.7%) children who were overweight and 478 (11.3%; 10.3-12.2%) children with obesity. Children with obesity (n=71; 14.9%) more commonly underwent (adeno)tonsillectomy than children of healthy weight (n=282; 9.3%; P<0.001; odds ratio [OR] 2.15; 95% CI: 1.58-2.92). Fewer children with obesity (n=365; 77% vs n=2552; 85%) were anaesthetised by consultant anaesthetists (OR 0.62; 95% CI: 0.48-0.79). Mask ventilation was difficult for 3.7% of children with obesity vs 0.6% of children of healthy weight (difference 3.0%; 95% CI: 1.3-4.7%; P<0.001). Comparison with NCMP data demonstrated an over-representation of obesity amongst the paediatric surgical population. CONCLUSIONS: This large multicentre cohort study suggests a concerning prevalence of children with obesity presenting for anaesthesia. These results should be used to inform optimal provision of care for this population and support perioperative healthcare initiatives to address the burden of childhood obesity. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03994419.

Women's perspectives on smartphone apps for fertility tracking and predicting conception: a mixed methods study.

BACKGROUND: Women use fertility tracking apps (FTAs) for conception purposes, but user perspectives on FTA use for conception are largely unknown. In collaboration with SPD Clearblue, this study explored: how women trying to conceive use FTAs; women's knowledge of their conception chances; and women's feelings towards a potential natural conception prediction app (NCPA). METHODS: A mixed methods design was used (online survey and phone interviews). Participants were women 18-40 years old actively trying to conceive. RESULTS: The survey received 154 responses and 24 interviews were conducted. Thematic analysis of interviews found that women consider several factors before trying to conceive (ex. age, financial and job security, stability of relationship, etc.) and may adopt lifestyle and behaviour changes when trying (ex. increasing exercise, smoking cessation, diet changes, etc.). Survey results indicated that nearly all respondents were aware of FTAs (n = 146, 94.8%), however, several other fertility and conception information sources were also used (ex. health care providers, online sources, family and friends, etc.). Nearly all respondents reported they would use an NCPA (n = 153, 99.4%). During interviews women had positive feelings towards such an app due to it offering new and individualised information, but worried the app could provide upsetting information. CONCLUSION: This research elaborates on women's uses of and interest in FTAs. Stakeholders should use this research to reflect on current conception experiences and possibilities for improvement through development of an NCPA. Future research should seek opinions from a more diverse sample of women to inform the development of an inclusive NCPA.

Predicting kidney failure risk after acute kidney injury among people receiving nephrology clinic care.

BACKGROUND: Outcomes after acute kidney injury (AKI) are well described, but not for those already under nephrology clinic care. This is where discussions about kidney failure risk are commonplace. We evaluated whether the established kidney failure risk equation (KFRE) should account for previous AKI episodes when used in this setting. METHODS: This observational cohort study included 7491 people referred for nephrology clinic care in British Columbia in 2003-09 followed to 2016. Predictors were previous Kidney Disease: Improving Global Outcomes-based AKI, age, sex, proteinuria, estimated glomerular filtration rate (eGFR) and renal diagnosis. Outcomes were 5-year kidney failure and death. We developed cause-specific Cox models (AKI versus no AKI) for kidney failure and death, stratified by eGFR (

Incidence of paediatric unplanned day-case admissions in the UK and Ireland: a prospective multicentre observational study.

BACKGROUND: Failure to discharge home after day-case procedures has a negative impact on patients, families, and hospital finances. There are currently no national paediatric data on the incidence and causes of unplanned admission. We determined the incidence of unplanned admissions after paediatric day-case anaesthesia, and identified risk factors leading to unplanned admission. METHODS: During a 6 week period (in October and November 2017), all children aged 16 yr or under receiving general anaesthesia without an inpatient bed on arrival were included. Hospital, surgical, and procedural details; anonymised patient characteristic data; and anaesthetic and surgical experience were collected by local Paediatric Anaesthesia Trainee Research Network coordinators. A mixed-effects binary logistic regression model with backward selection was used to determine variables associated with unplanned admission. RESULTS: Ninety three hospitals across the UK and Ireland participated. There were 25 986 cases, of which 640 were unplanned admissions. The independent risk factors for unplanned admission were ASA-physical status (PS) (ASA-PS 3/4 vs ASA-PS 1; odds ratio [OR]: 2.80 [95% confidence interval {CI}: 2.07-3.77]), duration of procedure (OR: 1.04 [95% CI: 1.03-1.05]), and surgical specialty (vs ear, nose, and throat [highest caseload specialty]: cardiology OR: 1.89 [95% CI: 1.15-3.06], orthopaedics/trauma OR: 0.91 [95% CI: 0.69-1.18], and general surgery OR: 0.59 [95% CI: 0.46-0.77]). The commonest reasons for admission were unexpected surgical complexity, pain, postoperative nausea and vomiting, and late finish. CONCLUSIONS: Paediatric patient physical status, some types of surgery and duration of procedure were associated with unplanned day-surgery admissions. Unexpected surgical complexity and patient discomfort in recovery were common factors.

A comparison of the beta-geometric model with landmarking for dynamic prediction of time to pregnancy.

We conducted a simulation study to compare two methods that have been recently used in clinical literature for the dynamic prediction of time to pregnancy. The first is landmarking, a semi-parametric method where predictions are updated as time progresses using the patient subset still at risk at that time point. The second is the beta-geometric model that updates predictions over time from a parametric model estimated on all data and is specific to applications with a discrete time to event outcome. The beta-geometric model introduces unobserved heterogeneity by modelling the chance of an event per discrete time unit according to a beta distribution. Due to selection of patients with lower chances as time progresses, the predicted probability of an event decreases over time. Both methods were recently used to develop models predicting the chance to conceive naturally. The advantages, disadvantages and accuracy of these two methods are unknown. We simulated time-to-pregnancy data according to different scenarios. We then compared the two methods by the following out-of-sample metrics: bias and root mean squared error in the average prediction, root mean squared error in individual predictions, Brier score and c statistic. We consider different scenarios including data-generating mechanisms for which the models are misspecified. We applied the two methods on a clinical dataset comprising 4999 couples. Finally, we discuss the pros and cons of the two methods based on our results and present recommendations for use of either of the methods in different settings and (effective) sample sizes.

Individual and Combined Impact of Heart Failure and Atrial Fibrillation on Ischemic Stroke Outcomes.

Background and Purpose- We aimed to determine individual and combined effects of atrial fibrillation (AF) and heart failure (HF) on acute ischemic stroke outcomes: in-hospital mortality, length-of-stay, and poststroke disability; long-term mortality and stroke recurrence. Methods- Prospective cohort study of patients with acute ischemic stroke admitted to a UK center with a catchment population of ≈900 000 between 2004 and 2016. Exposure groups were patients with neither AF nor HF (reference group), those with AF but without HF, those with HF but without AF, and those with AF+HF. Logistic and Cox regressions were used to model in-hospital and long-term outcomes, respectively. Results- A total of 10 816 patients with a mean age±SD =77.9±12.1 years, 48% male were included. Only 30 (4.9%) of the patients with HF but not AF were anticoagulated at discharge. Both AF (odds ratio, 1.24 [95% CI, 1.07-1.43]), HF (odds ratio, 1.40 [1.10-1.79]), and their combination (odds ratio, 2.23 [1.83-2.72]) were associated with increased odds of in-hospital mortality. All 3 exposure groups were associated with increased length-of-stay, while only AF predicted increased disability (1.36 [1.12-1.64]). Patients were followed for a median of 5.5 and 3.7 years for mortality and recurrence, respectively. Long-term mortality was associated with AF (hazard ratio, 1.45 [95% CI, 1.33-1.59]), HF (2.07 [1.83-2.36]), and their combination (2.20 [1.96-2.46]). Recurrent stroke was associated with AF 1.50 (1.26-1.78), HF (1.33 [1.01-1.75]), and AF with HF (1.62 [1.28-2.07]). Conclusions- The AF-associated excess risk of stroke recurrence was independent of comorbid HF. HF without AF was also associated with a significant risk of recurrence. Anticoagulation for secondary stroke prevention in patients with HF without AF may require further evaluation in a clinical trial setting.

Calibration: the Achilles heel of predictive analytics.

BACKGROUND: The assessment of calibration performance of risk prediction models based on regression or more flexible machine learning algorithms receives little attention. MAIN TEXT: Herein, we argue that this needs to change immediately because poorly calibrated algorithms can be misleading and potentially harmful for clinical decision-making. We summarize how to avoid poor calibration at algorithm development and how to assess calibration at algorithm validation, emphasizing balance between model complexity and the available sample size. At external validation, calibration curves require sufficiently large samples. Algorithm updating should be considered for appropriate support of clinical practice. CONCLUSION: Efforts are required to avoid poor calibration when developing prediction models, to evaluate calibration when validating models, and to update models when indicated. The ultimate aim is to optimize the utility of predictive analytics for shared decision-making and patient counseling.

Three myths about risk thresholds for prediction models.

BACKGROUND: Clinical prediction models are useful in estimating a patient's risk of having a certain disease or experiencing an event in the future based on their current characteristics. Defining an appropriate risk threshold to recommend intervention is a key challenge in bringing a risk prediction model to clinical application; such risk thresholds are often defined in an ad hoc way. This is problematic because tacitly assumed costs of false positive and false negative classifications may not be clinically sensible. For example, when choosing the risk threshold that maximizes the proportion of patients correctly classified, false positives and false negatives are assumed equally costly. Furthermore, small to moderate sample sizes may lead to unstable optimal thresholds, which requires a particularly cautious interpretation of results. MAIN TEXT: We discuss how three common myths about risk thresholds often lead to inappropriate risk stratification of patients. First, we point out the contexts of counseling and shared decision-making in which a continuous risk estimate is more useful than risk stratification. Second, we argue that threshold selection should reflect the consequences of the decisions made following risk stratification. Third, we emphasize that there is usually no universally optimal threshold but rather that a plausible risk threshold depends on the clinical context. Consequently, we recommend to present results for multiple risk thresholds when developing or validating a prediction model. CONCLUSION: Bearing in mind these three considerations can avoid inappropriate allocation (and non-allocation) of interventions. Using discriminating and well-calibrated models will generate better clinical outcomes if context-dependent thresholds are used.

Treatment-independent live birth after in-vitro fertilisation: a retrospective cohort study of 2,133 women.

STUDY QUESTION: What is the chance of a treatment-independent live birth following IVF (including ICSI) treatment? SUMMARY ANSWER: Over 5 years of follow-up, the treatment-independent live birth rate was 17% in unsuccessfully treated women and 15% in those who had a live birth after IVF. WHAT IS KNOWN ALREADY: A limited number of studies have investigated the chance of treatment-independent conception following completion of IVF, but most of them have been based on surveys with poor response rates and limited sample sizes. STUDY DESIGN, SIZE, DURATION: This is a population-based, retrospective cohort study of 2133 women who received IVF treatment between 1998 and 2011 at a single regional IVF Unit and were followed for a minimum of 1 year and maximum of 15 years after their last IVF or ICSI treatment cycle. PARTICIPANTS/MATERIALS, SETTING, METHODS: This study included all women, residing in the north-east of the UK, who attended the Aberdeen Fertility Clinic and received IVF treatment between 1998 and 2011. Clinical and diagnostic information of all women was linked with treatment and pregnancy outcome data. A total of 2133 women were divided into two groups: (i) those who achieved a live birth following successful IVF or ICSI treatment (n = 1060) and (ii) those in whom treatment was unsuccessful i.e. resulted in either no pregnancy or pregnancy loss (n = 1073). The two groups were followed from the date of the last embryo transfer until the first treatment-independent live birth or 31 December 2012, whichever came first. The primary outcome was the treatment-independent live birth rate at 1, 2.5, 5 and 10 years of follow-up. Cox regression was used to determine factors associated with treatment-independent live birth in each group. MAIN RESULTS AND THE ROLE OF CHANCE: Within 5 years of follow-up, the treatment-independent live birth rate was 17% (95% CI, 15-19%) among women whose IVF or ICSI treatment was unsuccessful and 15% (95% CI, 12-17%) among women whose treatment resulted in live birth. In both groups, shorter duration of infertility, younger female age and IVF as compared to ICSI were associated with a higher chance of achieving treatment-independent live birth. Among unsuccessfully treated women, the chance of post-IVF live birth was reduced in those with tubal factor infertility. Three or more previous IVF or ICSI embryo transfers were associated with a lower chance of treatment-independent live birth among successfully treated women. LIMITATIONS, REASONS FOR CAUTION: The study was conducted in a single fertility centre, which could compromise the generalizability of the findings. Moreover, data were unavailable on the women's use of contraception or active attempts to get pregnant, both of which could influence treatment-independent live birth rates. WIDER IMPLICATIONS OF THE FINDINGS: This study provides a better understanding of the long-term prognosis for treatment-independent live birth after completion of IVF or ICSI treatment. The results will inform women of their chances of a treatment-independent live birth following failed or successful treatment and the factors that are associated with it. STUDY FUNDING/COMPETING INTEREST(S): This work was funded by a Chief Scientist Office Postdoctoral Training Fellowship in Health Services Research and Health of the Public Research (Ref PDF/12/06). The views expressed here are those of the authors and not necessarily those of the Chief Scientist Office. The authors have no competing interests. TRIAL REGISTRATION NUMBER: Not applicable.

Myocardial infarction after acute ischaemic stroke: Incidence, mortality and risk factors.

OBJECTIVES: To determine the risk factor profiles associated with post-acute ischaemic stroke (AIS) myocardial infarction (MI) over long-term follow-up. METHODS: This observational study includes prospectively identified AIS patients (n = 9840) admitted to a UK regional centre between January 2003 and December 2016 (median follow-up: 4.72 years). Predictors of post-stroke MI during follow-up were examined using logistic and Cox regression models for in-hospital and post-discharge events, respectively. MI incidence was determined using a competing risk non-parametric estimator. The influence of post-stroke MI on mortality was examined using Cox regressions. RESULTS: Mean age (SD) of study participants was 77.3 (12.2) years (48% males). Factors associated with in-hospital MI (OR [95% CI]) were increasing blood glucose (1.80 [1.17-2.77] per 10 mmol/L), total leucocyte count (1.25 [1.01-1.54] per 10 × 109 /L) and CRP (1.05 [1.02-1.08] per 10 mg/L increase). Age (HR [95% CI] = 1.03 [1.01-1.06]), coronary heart disease (1.59 [1.01-2.50]), chronic kidney disease (2.58 [1.44-4.63]) and cancers (1.76 [1.08-2.89]) were associated with incident MI between discharge and one-year follow-up. Age (1.02 [1.00-1.03]), diabetes (1.96 [1.38-2.65]), congestive heart failure (2.07 [1.44-2.99]), coronary heart disease (1.81 [1.31-2.50]), hypertension [1.86 (1.24-2.79)] and peripheral vascular disease (2.25 [1.40-3.63]) were associated with incident MI between 1 and 5 years after discharge. Diabetes (2.01 [1.09-3.72]), hypertension (3.69 [1.44-9.45]) and peripheral vascular disease (2.46 [1.02-5.98]) were associated with incident MI between 5 and 10 years after discharge. Cumulative MI incidence over 10 years was 5.4%. MI during all follow-up periods (discharge-1, 1-5, 5-10 years) was associated with increased risk of death (respective HR [95% CI] = 3.26 [2.51-4.15], 1.96 [1.58-2.42] and 1.92 [1.26-2.93]). CONCLUSIONS: In conclusion, prognosis is poor in post-stroke MI. We highlight a range of potential areas to focus preventative efforts.

Post-discharge kidney function is associated with subsequent ten-year renal progression risk among survivors of acute kidney injury.

The extent to which renal progression after acute kidney injury (AKI) arises from an initial step drop in kidney function (incomplete recovery), or from a long-term trajectory of subsequent decline, is unclear. This makes it challenging to plan or time post-discharge follow-up. This study of 14651 hospital survivors in 2003 (1966 with AKI, 12685 no AKI) separates incomplete recovery from subsequent renal decline by using the post-discharge estimated glomerular filtration rate (eGFR) rather than the pre-admission as a new reference point for determining subsequent renal outcomes. Outcomes were sustained 30% renal decline and de novo CKD stage 4, followed from 2003-2013. Death was a competing risk. Overall, death was more common than subsequent renal decline (37.5% vs 11.3%) and CKD stage 4 (4.5%). Overall, 25.7% of AKI patients had non-recovery. Subsequent renal decline was greater after AKI (vs no AKI) (14.8% vs 10.8%). Renal decline after AKI (vs no AKI) was greatest among those with higher post-discharge eGFRs with multivariable hazard ratios of 2.29 (1.88-2.78); 1.50 (1.13-2.00); 0.94 (0.68-1.32) and 0.95 (0.64-1.41) at eGFRs of 60 or more; 45-59; 30-44 and under 30, respectively. The excess risk after AKI persisted over ten years of study, irrespective of AKI severity, or post-episode proteinuria. Thus, even if post-discharge kidney function returns to normal, hospital admission with AKI is associated with increased renal progression that persists for up to ten years. Follow-up plans should avoid false reassurance when eGFR after AKI returns to normal.

Cumulative live birth rates following miscarriage in an initial complete cycle of IVF: a retrospective cohort study of 112 549 women.

STUDY QUESTION: In women undergoing IVF/ICSI who miscarry in their first complete cycle, what is the chance of a live birth in subsequent complete cycles, and how does this compare with those whose first complete cycle ends with live birth or without a pregnancy? SUMMARY ANSWER: After two further complete cycles of IVF/ICSI, women who had miscarried or had a live birth in their first complete cycle had a higher chance of live birth (40.9 and 49.0%, respectively) than those who had no pregnancies (30.1%). WHAT IS KNOWN ALREADY: Cumulative live birth rates (CLBRs) after one or more complete cycles of IVF have been reported previously, as have some of the risk factors associated with miscarriage, both in general populations and in those undergoing IVF. Chances of cumulative live birth after a number of complete IVF cycles involving replacement of fresh followed by frozen embryos after an initial miscarriage in a population undergoing IVF treatment have not been reported previously. STUDY DESIGN, SIZE, DURATION: National population-based cohort study of 112 549 women who started their first IVF treatment between 1999 and 2008. PARTICIPANTS/MATERIALS, SETTING, METHODS: Data from the UK Human Fertilisation and Embryology Authority (HFEA) register on IVF/ICSI treatments, using autologous gametes were analysed. CLBRs were estimated in women who (i) had miscarriage (and no live birth), (ii) at least one live birth or (iii) no pregnancy in their first complete cycle of IVF/ICSI (including fresh and frozen embryo transfers following a single oocyte retrieval episode). A multivariable analysis was performed to assess the effect of first complete cycle outcome on subsequent CLBRs after adjusting for confounding factors such as female age, duration of infertility and cause of infertility. MAIN RESULTS AND THE ROLE OF CHANCE: In their first complete cycle, 9321 (8.3%) women had at least one miscarriage (and no live birth); 33 152 (29.5%) had at least one live birth and 70 076 (62.3%) had no pregnancies. After two further complete cycles, conservative CLBRs (which assume that women who discontinued treatment subsequently never had a live birth) were 40.9, 49.0 and 30.1%, while optimal CLBRs (which assume that women who discontinue have the same chance of live birth as those treated) were 49.5, 57.9 and 38.4% in the miscarriage, live birth and no pregnancy groups respectively. Odds of cumulative live birth for women who miscarried in their first complete cycle were 42% higher than those who had no pregnancy [odds ratio (95% CI) = 1.42 (1.34, 1.50)], and twice as high for live birth versus no pregnancy [2.04 (1.89, 2.20)]. Negative predictors for live birth in all women included tubal infertility [0.88 (0.82, 0.94)] and increasing age [18-40 years = 0.94 (0.94, 0.95); >40 years = 0.63 (0.59, 0.66) per year]. LIMITATIONS AND REASON FOR CAUTION: CLBRs could not be estimated for treatments occurring after September 2008 due to potentially incomplete data following regulatory changes regarding consent for data use in research. Additionally, covariates not included in the HFEA database (including BMI, smoking, previous history of miscarriage and gestational age at miscarriage) could not be adjusted for in our analysis. WIDER IMPLICATIONS OF THE FINDINGS: Miscarriage following IVF can be devastating for couples who are uncertain about their ultimate prognosis. Our findings will provide reassurance to these couples as they consider their options for continuing treatment. STUDY FUNDING/COMPETING INTEREST(S): N.J.C. received an Aberdeen Summer Research Scholarship funded by the Institute of Applied Health Sciences (University of Aberdeen), through the Aberdeen Clinical Academic Training Scheme. This work was supported by a Chief Scientist Office Postdoctoral Training Fellowship in Health Services Research and Health of the Public Research (Ref PDF/12/06). The views expressed here are those of the authors and not necessarily those of the Chief Scientist Office or the University of Aberdeen. The funders did not have any role in the study design; in the collection, analysis and interpretation of data; in the writing of the report; nor in the decision to submit the paper for publication. None of the authors has any conflicts of interest to declare.

Acute kidney injury as an independent risk factor for unplanned 90-day hospital readmissions.

BACKGROUND: Reducing readmissions is an international priority in healthcare. Acute kidney injury (AKI) is common, serious and also a global concern. This analysis evaluates AKI as a candidate risk factor for unplanned readmissions and determines the reasons for readmissions. METHODS: GLOMMS-II is a large population cohort from one health authority in Scotland, combining hospital episode data and complete serial biochemistry results through data-linkage. 16453 people (2623 with AKI and 13830 without AKI) from GLOMMS-II who survived an index hospital admission in 2003 were used to identify the causes of and predict readmissions. The main outcome was "unplanned readmission or death" within 90 days of discharge. In a secondary analysis, the outcome was limited to readmissions with acute pulmonary oedema. 26 candidate predictors during the index admission included AKI (defined and staged 1-3 using an automated e-alert algorithm), prior AKI episodes, baseline kidney function, index admission circumstances and comorbidities. Prediction models were developed and assessed using multivariable logistic regression (stepwise variable selection), C statistics, bootstrap validation and decision curve analysis. RESULTS: Three thousand sixty-five (18.6%) patients had the main outcome (2702 readmitted, 363 died without readmission). The outcome was strongly predicted by AKI. Multivariable odds ratios for AKI stage 3; 2 and 1 (vs no AKI) were 2.80 (2.22-3.53); 2.23 (1.85-2.68) and 1.50 (1.33-1.70). Acute pulmonary oedema was the reason for readmission in 26.6% with AKI and eGFR < 60; and 4.0% with no AKI and eGFR ≥ 60. The best stepwise model from all candidate predictors had a C statistic of 0.698 for the main outcome. In a secondary analysis, a model for readmission with acute pulmonary oedema had a C statistic of 0.853. In decision curve analysis, AKI improved clinical utility when added to any model, although the incremental benefit was small when predicting the main outcome. CONCLUSIONS: AKI is a strong, consistent and independent risk factor for unplanned readmissions - particularly readmissions with acute pulmonary oedema. Pre-emptive planning at discharge should be considered to minimise avoidable readmissions in this high risk group.

Planned Repeat Cesarean Section at Term and Adverse Childhood Health Outcomes: A Record-Linkage Study.

BACKGROUND: Global cesarean section (CS) rates range from 1% to 52%, with a previous CS being the commonest indication. Labour following a previous CS carries risk of scar rupture, with potential for offspring hypoxic brain injury, leading to high rates of repeat elective CS. However, the effect of delivery by CS on long-term outcomes in children is unclear. Increasing evidence suggests that in avoiding exposure to maternal bowel flora during labour or vaginal birth, offspring delivered by CS may be adversely affected in terms of energy uptake from the gut and immune development, increasing obesity and asthma risks, respectively. This study aimed to address the evidence gap on long-term childhood outcomes following repeat CS by comparing adverse childhood health outcomes after (1) planned repeat CS and (2) unscheduled repeat CS with those that follow vaginal birth after CS (VBAC). METHODS AND FINDINGS: A data-linkage cohort study was performed. All second-born, term, singleton offspring delivered between 1 January 1993 and 31 December 2007 in Scotland, UK, to women with a history of CS (n = 40,145) were followed up until 31 January 2015. Outcomes assessed included obesity at age 5 y, hospitalisation with asthma, learning disability, cerebral palsy, and death. Cox regression and binary logistic regression were used as appropriate to compare outcomes following planned repeat CS (n = 17,919) and unscheduled repeat CS (n = 8,847) with those following VBAC (n = 13,379). Risk of hospitalisation with asthma was greater following both unscheduled repeat CS (3.7% versus 3.3%, adjusted hazard ratio [HR] 1.18, 95% CI 1.05-1.33) and planned repeat CS (3.6% versus 3.3%, adjusted HR 1.24, 95% CI 1.09-1.42) compared with VBAC. Learning disability and death were more common following unscheduled repeat CS compared with VBAC (3.7% versus 2.3%, adjusted odds ratio 1.64, 95% CI 1.17-2.29, and 0.5% versus 0.4%, adjusted HR 1.50, 95% CI 1.00-2.25, respectively). Risk of obesity at age 5 y and risk of cerebral palsy were similar between planned repeat CS or unscheduled repeat CS and VBAC. Study limitations include the risk that women undergoing an unscheduled CS had intended to have a planned CS, and lack of data on indication for CS, which may confound the findings. CONCLUSIONS: Birth by repeat CS, whether planned or unscheduled, was associated with an increased risk of hospitalisation with asthma but no difference in risk of obesity at age 5 y. Greater risk of death and learning disability following unscheduled repeat CS compared to VBAC may reflect complications during labour. Further research, including meta-analyses of studies of rarer outcomes (e.g., cerebral palsy), are needed to confirm whether such risks are similar between delivery groups.