Evidence based policy decisions through a Bayesian approach: The case of a statin appraisal in the Netherlands

Child Abuse, Misdiagnosed by an Expertise Center-Part II-Misuse of Bayes' Theorem

A newborn girl had, from two weeks on, small bruises on varying body locations, but not on her chest. Her Armenian grandmother easily bruised, too. Her mother was diagnosed with hypermobility-type Ehlers-Danlos-Syndrome (hEDS), an autosomal dominant connective tissue disorder, with a 50% inheritance probability. Referral to a University Medical Center located "Dutch Expertise Center for Child Abuse" resulted (prior to consultation) in physical abuse suspicion. Protocol-based skeletal X-rays showed three healed, asymptomatic rib fractures. A protocol-based Bayesian likelihood ratio guesstimation gave 10-100, erroneously used to suggest a 10-100 times likelier non-accidental-than-accidental cause. Foster care placement followed, even in a secret home, where she also bruised, suggesting hEDS inheritance. Correct non-accidental/accidental Bayes' probability of symptoms is (likelihood ratio) × (physical abuse incidence). From the literature, we derived an infant abuse incidence between about ≈0.0009 and ≈0.0026 and a likelihood ratio of <5 for bruises. For rib fractures, we used a zero likelihood ratio, arguing their cause was birth trauma from the extra delivery pressure on the chest, combined with fragile bones as the daughter of an hEDS-mother. We thus derived a negligible abuse/accidental probability between <5 × 0.0009 <0.005 and <5 × 0.0026 <0.013. The small abuse incidence implies that correctly using Bayes' theorem will also miss true infant physical abuse cases. Curiously, because likelihood ratios assess how more often symptoms develop if abuse did occur versus non-abuse, Bayes' theorem then implies a 100% infant abuse incidence (unwittingly) used by LECK. In conclusion, probabilities should never replace differential diagnostic procedures, the accepted medical method of care. Well-known from literature, supported by the present case, is that (child abuse pediatrics) physicians, child protection workers, and judges were unlikely to understand Bayesian statistics. Its use without statistics consultation should therefore not have occurred. Thus, Bayesian statistics, and certainly (misused) likelihood ratios, should never be applied in cases of physical child abuse suspicion. Finally, parental innocence follows from clarifying what could have caused the girl's bruises (inherited hEDS), and rib fractures (birth trauma from fragile bones).

Looking beyond the forest: Using harvest plots, gap analysis, and expert consultations to assess effectiveness, engage stakeholders, and inform policy

We describe a combination of methods for assessing the effectiveness of complex interventions, especially where substantial heterogeneity with regard to the population, intervention, comparison, outcomes, and study design of interest is expected. We applied these methods in a recent systematic review of the effectiveness of reinforced home-based palliative care (rHBPC) interventions, which included home-based care with an additional and explicit component of lay caregiver support. We first summarized the identified evidence, deemed inappropriate for statistical pooling, graphically by creating harvest plots. Although very useful as a tool for summary and presentation of overall effectiveness, such graphical summary approaches may obscure relevant differences between studies. Thus, we then used a gap analysis and conducted expert consultations to look beyond the aggregate level at how the identified evidence of effectiveness may be explained. The goal of these supplemental methods was to step outside of the conventional systematic review and explore this heterogeneity from a broader perspective, based on the experience of palliative care researchers and practitioners. The gap analysis and expert consultations provided valuable input into possible underlying explanations in the evidence, which could be helpful in the further adaptation and testing of existing rHBPC interventions or the development and evaluation of new ones. We feel that such a combination of methods could prove accessible, understandable, and useful in informing decisions and could thus help increase the relevance of systematic reviews to the decision-making process.

Prevalence of Neuropathic Pain in Cancer Patients: Pooled Estimates From a Systematic Review of Published Literature and Results From a Survey Conducted in 50 Italian Palliative Care Centers

CONTEXT

Because of the increasing body of literature on neuropathic cancer pain (NCP), an accurate estimate of its prevalence requires recurring updates.

OBJECTIVES

To provide this estimate using information from a systematic review and a survey.

METHODS

Using MEDLINE, Embase, and a previous review, we searched for studies published up to 2014 reporting data on NCP prevalence in adult cancer populations. Pooled prevalence rates from observational prospective studies were computed. The association between NCP prevalence and possible predictors was investigated for oncology and palliative settings. Prevalence rates were extracted from a questionnaire answered by 137 physicians working in 50 Italian centers of palliative care. Estimates from studies conducted in palliative settings and from the experts were analyzed separately and eventually pooled with an informative Bayesian random-effect model.

RESULTS

Twenty-nine observational studies were identified. The overall pooled prevalence was 31.2%, with high heterogeneity; similar figures were observed when oncology and palliative settings were individually considered. A slightly higher prevalence of NCP was detected for hospice/inpatients as compared to outpatients, in both settings. The mean NCP prevalence reported by the survey experts was 44.2%; the pooled Bayesian estimate for the palliative setting corresponded to 43.0% (95% CI: 40.0-46.0). The subgroup with the lowest heterogeneity and where the literature and experts' estimates were closest is hospice/inpatients, with a pooled Bayesian prevalence rate of 34.9% (95% CI: 29.9-41.0).

CONCLUSION

The systematic review and the survey suggest that more than one in three patients with cancer pain also experiences NCP.

Synthesis of evidence for reimbursement decisions: a Bayesian reanalysis

OBJECTIVES

The aim of this study was to compare Bayesian methods with the standard methods that are used for evidence-based policy making.

METHODS

We performed a Bayesian reanalysis of the data underlying a reimbursement advice by the Dutch National Health Insurance Board (CVZ) regarding the anti-diabetic drug exenatide (an alternative to insulin). We synthesized evidence from various sources that was available when the CVZ advice was drafted: expert opinion (as elicited from internists), experimental data (from direct comparison studies), and observational data. Subsequently, the original frequentist results and the results from the Bayesian reanalysis were compared in terms of outcomes and interpretations. These results were presented in a meeting with staff from CVZ, whose opinions about the usefulness of a Bayesian approach were assessed using a questionnaire.

RESULTS

The Bayesian approach yields outcomes that summarize different pieces of evidence, which would have been difficult to obtain otherwise. Moreover, there are conceptual differences, and the Bayesian approach allows for determining probabilities of clinically relevant differences. The staff at CVZ were fairly positive with respect to the use of Bayesian methods, although practical barriers were also seen as important.

CONCLUSIONS

The Bayesian outcomes are different and could be more suited to the informational needs of policy makers. The response from staff at CVZ provides some support for this statement, but more research at the interface of science and policy is needed.