Using alternative statistical formats for presenting risks and risk reductions

 

Abstract

Background

The success of evidence-based practice depends on the clear and effective communication of statistical information.

Objectives

To evaluate the effects of using alternative statistical presentations of the same risks and risk reductions on understanding, perception, persuasiveness and behaviour of health professionals, policy makers, and consumers.

Search methods

We searched Ovid MEDLINE (1966 to October 2007), EMBASE (1980 to October 2007), PsycLIT (1887 to October 2007), and the Cochrane Central Register of Controlled Trials (The Cochrane Library, 2007, Issue 3). We reviewed the reference lists of relevant articles, and contacted experts in the field.

Selection criteria

We included randomized and non-randomized controlled parallel and cross-over studies. We focused on four comparisons: a comparison of statistical presentations of a risk (eg frequencies versus percentages) and three comparisons of statistical presentation of risk reduction: relative risk reduction (RRR) versus absolute risk reduction (ARR), RRR versus number needed to treat (NNT), and ARR versus NNT.

Data collection and analysis

Two authors independently selected studies for inclusion, extracted data, and assessed risk of bias. We contacted investigators to obtain missing information. We graded the quality of evidence for each outcome using the GRADE approach. We standardized the outcome effects using adjusted standardized mean difference (SMD).

Main results

We included 35 studies reporting 83 comparisons. None of the studies involved policy makers. Studies of alternative formats for presenting risks focused on either diagnostic or screening tests. Participants (health professionals and consumers) understood natural frequencies better than percentages (SMD 0.69 (95% confidence interval (CI) 0.45 to 0.93)). In studies of alternative formats for presenting risk reductions of interventions, and compared with ARR, RRR had little or no difference in understanding (SMD 0.02 (95% CI -0.39 to 0.43)) but was perceived to be larger (SMD 0.41 (95% CI 0.03 to 0.79)) and more persuasive (SMD 0.66 (95% CI 0.51 to 0.81)). Compared with NNT, RRR was better understood (SMD 0.73 (95% CI 0.43 to 1.04)), was perceived to be larger (SMD 1.15 (95% CI 0.80 to 1.50)) and was more persuasive (SMD 0.65 (95% CI 0.51 to 0.80)). Compared with NNT, ARR was better understood (SMD 0.42 (95% CI 0.12 to 0.71)), was perceived to be larger (SMD 0.79 (95% CI 0.43 to 1.15)).There was little or no difference for persuasiveness (SMD 0.05 (95% CI -0.04 to 0.15)). The sensitivity analyses including only high quality comparisons showed consistent results for persuasiveness for all three comparisons. Overall there were no differences between health professionals and consumers. The overall quality of evidence was rated down to moderate because of the use of surrogate outcomes and/or heterogeneity. None of the comparisons assessed behaviour.

Authors' conclusions

Natural frequencies are probably better understood than percentages in the context of diagnostic or screening tests. For communicating risk reductions, relative risk reduction (RRR), compared with absolute risk reduction (ARR) and number needed to treat (NNT), may be perceived to be larger and is more likely to be persuasive. However, it is uncertain whether presenting RRR is likely to help people make decisions most consistent with their own values and, in fact, it could lead to misinterpretation. More research is needed to further explore this question.

 

Plain language summary

Using different statistical formats for presenting health information

Examples illustrating the statistical terms used in this summary:

You read that a study found that an osteoporosis drug cuts the risk of having a hip fracture in the next three years by 50%.  Specifically, 10% of the untreated people had a hip fracture at three years, compared with 5% of the people who took the osteoporosis drug every day for three years.  Thus 5% (10% minus 5%) less people would suffer a hip fracture if they take the drug for 3 years. In other words, 20 patients need to take the osteoporosis drug over 3 years for an additional patient to avoid a hip fracture. "Cuts the risk of fracture by 50%" represents a relative risk reduction. "Five per cent less would suffer a fracture" represents an absolute risk reduction. "Twenty patients need to take the osteoporosis drug over 3 years for an additional patient to avoid a hip fracture" represents a number needed to treat.

You read that another study found that the risk of suffering a hip fracture over a three year period among people not taking any osteoporotic drug is 10%; another way of expressing this risk would be: 100 of 1000 people not taking any osteoporotic drug will suffer a hip fracture over a three year period. "10%" represents a percentage while "100 of 1000" represents a frequency.

Summary:

Health professionals and consumers may change their choices when the same risks and risk reductions are presented using alternative statistical formats. Based on the results of 35 studies reporting 83 comparisons, we found the risk of a health outcome is better understood when it is presented as a natural frequency rather than a percentage for diagnostic and screening tests. For interventions, and on average, people perceive risk reductions to be larger and are more persuaded to adopt a health intervention when its effect is presented in relative terms (eg using relative risk reduction which represents a proportional reduction) rather than in absolute terms (eg using absolute risk reduction which represents a simple difference). We found no differences between health professionals and consumers. The implications for clinical and public health practice are limited by the lack of research on how these alternative presentations affect actual behaviour. However, there are strong logical arguments for not reporting relative values alone, as they do not allow a fair comparison of benefits and harms as absolute values do.

Please refer to the Cochrane Collaboration Glossary for further explanations of the statistical terms used in this review.

 

Resumen

Antecedentes

Uso de formatos estadísticos alternativos para la presentación de riesgos y la reducción de riesgos

El éxito de la práctica basada en la evidencia depende de una comunicación clara y eficaz de la información estadística.

Objetivos

Evaluar los efectos del uso de presentaciones estadísticas alternativas de los mismos riesgos y las reducciones de riesgos para la comprensión, la percepción, la capacidad de persuasión y la conducta de los profesionales de la salud, los elaboradores de políticas y los usuarios.

Estrategia de búsqueda

Se hicieron búsquedas en Ovid MEDLINE (1966 hasta octubre 2007), EMBASE (1980 hasta octubre 2007), PsycLIT (1887 hasta octubre 2007), y en el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials) (The Cochrane Library, 2007, número 3). Se revisaron las listas de referencias de los artículos relevantes y se contactó con expertos en este tema.

Criterios de selección

Se incluyeron estudios aleatorios y no aleatorios controlados paralelos y cruzados. Nos centramos en cuatro comparaciones: una comparación de las presentaciones estadísticas de un riesgo (p.ej. frecuencias versus probabilidades) y tres comparaciones de la presentación estadística de reducción de riesgos: la reducción del riesgo relativo (RRR) versus reducción del riesgo absoluto (RRA), RRR versus número necesario a tratar (NNT) y RRA versus NNT.

Obtención y análisis de los datos

Dos autores seleccionaron de forma independiente los estudios para la inclusión, extrajeron los datos y evaluaron el riesgo de sesgo. Se estableció contacto con los investigadores para obtener la información que faltaba. Se calificó la calidad de las pruebas para cada resultado mediante el enfoque GRADE. Se estandarizaron los efectos de los resultados mediante el uso de la diferencia de medias estandarizada (DME) ajustada.

Resultados principales

Se incluyeron 35 estudios que informaban 83 comparaciones. Ninguno de los estudios incluyó elaboradores de políticas. Los participantes (profesionales de la salud y usuarios) comprendieron las frecuencias naturales mejor que las probabilidades (DME 0,69 [intervalo de confianza (IC) del 95%: 0,45 a 0,93]). Comparado con el RRA, el RRR presentó poca o ninguna diferencia en la comprensión (DME 0,02 [IC del 95%: −0,39 a 0,43]) pero se percibió de mayor tamaño (DME 0,41 [IC del 95%: 0,03 a 0,79]) y más convincente (DME 0,66 [IC del 95%: 0,51 a 0,81]). Comparado con el NNT, el RRR se comprendió mejor (DME 0,73 [IC del 95%: 0,43 a 1,04]), se percibió como más grande (DME 1,15 [IC del 95%: 0,80 a 1,50]) y fue más convincente (DME 0,65 [IC del 95%: 0,51 a 0,80]). Comparado con el NNT, el RRA se comprendió mejor (DME 0,42 [IC del 95%: 0,12 a 0,71]), se percibió como más grande (DME 0,79 [IC del 95%: 0,43 a 1.15]). Hubo poca o ninguna diferencia en la capacidad de persuasión (DME 0,05 [IC del 95%: −0,04 a 0,15]). Los análisis de sensibilidad que incluyeron sólo las comparaciones de alta calidad mostraron resultados coherentes de la capacidad de persuasión para las tres comparaciones. En general, no hubo diferencias entre los profesionales sanitarios y los usuarios. La calidad general de las pruebas se valoró como baja a intermedia debido al uso de resultados alternativos o a la heterogeneidad. Ninguna de las comparaciones evaluó el comportamiento.

Conclusiones de los autores

Las frecuencias naturales quizás se comprendan mejor que las probabilidades. La reducción del riesgo relativo (RRR), comparada con la reducción del riesgo absoluto (RRA) y el número necesario a tratar (NNT), puede percibirse como más grande y tiene mayor probabilidad de ser convincente. Sin embargo, no puede precisarse si la presentación de la RRR ayudaría a las personas a que tomen decisiones más compatibles con sus propios valores y, en realidad, podría llevar a una interpretación errónea. Se necesitan más estudios de investigación para explorar esta cuestión.

Traducción

Traducción realizada por el Centro Cochrane Iberoamericano