Author
Year
Country
|
Study
design
|
Disease
category and setting
|
Intervention and comparison
|
Significant results and conclusions
|
Comments and analysis
|
Fraccaro et al., 2018, UK (22).
|
Controlled trial with 20 patients.
|
Kidney transplanted patients viewing hypothetic laboratory test results /scenarios at an online patient portal.
|
Participants viewed three different graphical presentations (of 28 blood tests) representing a low, medium and high-risk clinical scenario.
Outcome: Accuracy of the participants’ interpretation of the risk, measured by three response options after each scenario:
Calling doctor immediately, arrange an appointment within four weeks, wait for next appointment within three months.
|
Findings were not significantly different. The study confirmed that the participants had difficulties when interpreting laboratory test results. Many participants (65%) underestimated the need for action at least once even when abnormal values were highlighted using colours and graphical cues.
|
Applicability
This study explored whether a visual presentation using colours and graphical cues can improve the patients’ ability to interpret the risk information presented.
Limitations:
Small cohort, limited statistical power.
No evaluation of graph literacy or numeracy at baseline.
The participants were all used to being monitored by biochemical tests and not comparable to the average population in general practice.
|
Valázquez-López et al., 2017, Mexico (23).
|
Randomised clinical trial with four primary care clinics and 351 patients.
|
Patients with type 2 diabetes (DM-2), without severe complications, in primary care.
|
Multimedia education program (MEP) and nutritional therapy (NT) compared to a control group who received NT only.
The NT was personalised according to comorbidities and nutritional preferences.
The NT + MEP group was educated through a MEP named Nutriluv®. A specific MEP module was shown in an informational kiosk prior to the nutritional session. Duration of intervention was 21 months.
|
Diabetes education with MEP was an effective strategy to improve the HbA1c (glycated haemoglobin), lipid profiles, and body weight in the patients with DM-2 in the long term.
|
Applicability
DM-2 is a common disease treated in primary care with potentially severe complication.
Limitations:
No statistical power calculation, weak statistical analysis e.g. conversion of units to percent, adjustments at baseline even though the groups were randomised.
No stratification of baseline characteristics.
Anthropometry measurements were not blinded.
The completion rates of patients were low (59.5% and 56.5%).
|
Perestelo-Pérez et al., 2016, Spain (24).
|
Cluster randomised trial with 29 doctors and 168 patients.
|
Cardiovascular disease (CVD) prevention in patients with DM-2 in primary care.
|
“Statin choice”, is an online clinical decision tool used in consultations in primary care. The decision aid calculates the risk of CVD in the next ten years, based on personal health information. The risk is displayed graphically with 100 dots coloured in green, red or yellow.
Evaluation of knowledge about statins, perception of CVD risk, decisional conflicts and satisfaction were assessed by questionnaires, immediately after the intervention and at follow up after three months.
Comparison: Usual care.
|
Intervention improved knowledge (p = 0.01), perception of the 10-year risk of myocardial infarction without using statins (p = 0.01) and satisfaction (p = 0.01).
The communication tool did not increase the length of consultations when compared with usual care.
The variance of consultation time was lower in the intervention group (p = 0.025), which suggests that the use of the decision tool may result in a more systematic and reproducible discussion. The decision tool improved the quality of the decision making about the use of statins.
|
Applicability
This decision tool and its outcome is relevant to risk communication in primary care.
Limitations:
No calculation of sample power.
Unbalanced randomisation regarding age, hypertension and number of patients taking statins at baseline.
Survey instruments were not checked for validity and reliability after the translation into Spanish.
Adherence after three months was self-reported and therefore may have been less reliable.
Doctors and patients were not blinded.
|
Peiris et al., 2015, Australia (25).
|
Randomised controlled trail (RCT) with 60 primary healthcare centres and 38725 patients.
|
Cardiovascular disease (CVD) risk management in primary healthcare.
|
A computer guided onscreen intervention in primary care. The intervention included a series of traffic light cues, to alert the general practitioner if the patient was not receiving sufficient screening or management. The intervention, for a minimum of 12 months, also included a graphical risk communication tool to assist the patient in understanding their CVD risk and how the risk could be affected by changes of individual risk factors.
Comparison: Usual care without the intervention tool or training of the general practitioner. Main outcomes were the fraction of patients receiving appropriate screening of risk factors and the proportion of patients receiving the recommended treatment according to guidelines.
|
The intervention was associated with improved measurements of the patients’ risk factors (62.8% vs. 53.4%, risk ratio 1.25 (95% CI, 1.04–1.50) and p = 0.02).
No significant differences in the proportions receiving guideline recommended medication prescriptions for the high-risk cohort (p = 0.12). There were significant treatment escalations for the high-risk cohort (new prescriptions or increased numbers of medicines). There was a higher proportion reaching guideline BP targets in the intervention group versus the control group. The intervention improved the CVD risk measurements and required minimal support.
|
Applicability
The pragmatic implementation of the tool was relevant to primary care. The outcome was clinical and a low level of implementation support was required.
Strengths:
Large sample size, power calculation has been made.
Adequate representativeness of the clinics included.
Sufficient randomisation with stratification.
Clinical outcome measures.
Outcome analysis were conducted blinded to randomisation.
Inclusion criteria were based on national guidelines for vascular screening.
Limitations:
The doctors received training as a part of the intervention.
Blinding of participants was not possible.
It was not possible to distinguish between the intervention’s effect on the patients and the practitioners risk understanding according to the study design.
|
Chmiel et al., 2014, Switzerland (26).
|
RCT with 30 general practices and 137 patients.
|
Patients with hypertension
(BP > 140 mmHg systolic and/or > 90 mmHg diastolic), treated in general practice.
|
Daily home BP measurement (HBPM) noted in either a schematic standard non-coloured BP booklet (control group) or a colour-coded booklet (intervention). The scheme in the coloured book was divided into three zones, according to the BP value: green, yellow and red. The duration of the study was six months. Clinical parameters and medication changes were recorded at 0, 3 and 6 months.
The outcome measurements: Adherence to HBPM measurements, BP values at follow up at the general practitioner and prescription of antihypertensive medication.
|
Findings showed no significant difference between the groups in absolute BP reduction or adherence with HBPM. The target BP (<140/90 mmHg) was achieved more often in the intervention group (43% vs. 25%; p = 0.044). No significant differences in adherence with HBPM, decrease in systolic and diastolic BP at end-point or change in Anti-hypertensive therapy (changed in 63 %)
|
Applicability
Simple, low cost and user friendly intervention with low necessity to understand numeracy. Study sample representing adult primary care patients with hypertension.
Calculation of statistical power and intention to treat analysis. Computer randomisation at patient level. Randomisation was adequate.
Precise and detailed manual for the HBPM in order to standardise outcome.
Limitations:
BP can be affected by medicine, exercise, stress, diet, lifestyle etc. The study design did not include guidelines or action plans according to the BP values. Therefore, the patients did not have a standardised way to respond if the BP was above normal value. Their response depended on their own beliefs and health literacy.
The majority of the patients (≈ 66 %) had already done HBPM before inclusion in the study.
The doctors and the patients were not blinded. Doctor and patient interaction was not investigated.
Possible Hawthorne effect. The calculated sample size was not attained, possible type 2 error.
|
Garcia-
Retamero et al., 2013, Spain (27).
|
Randomised trial with 81 general practitioners and 81 patients from four hospitals.
|
Questions regarding diagnostic inferences of cancer and diabetes.
|
Recruitment during an ordinary consultation and subsequent randomisation into four groups (as shown below).
Risk information given as:
|
Natural frequen-cies
|
Probabili-ties
|
Numerical
|
A
|
B
|
Numerical + visual tool
|
C
|
D
|
In addition, participants completed a numeracy test with 12 items. After receiving information about the prevalence of the disease, and the sensitivity and false-positive rate of the test for a given task, participants made the diagnostic inference about three medical tests.
The outcome measurements: Improvement in diagnostic inferences measured in probabilities or percentages of people having the disease.
Accuracy, perceived usefulness and perceived difficulty with the data representation were also assessed.
|
Performance was better when the information was presented in natural frequencies and presented both numerically and visually, as compared to probabilities and only numerical.
Visual tools improved the accuracy of diagnostic inference for medical doctors and their patients regardless of the numerical format.
Numerical format, visual aid, type of participant, level of numeracy as a covariate, and estimates of task difficulty as the only dependent variable, showed a main effect of the visual aid (p = 0.016)
The patients estimated information as less useful when it was provided only numerically, as compared to the same information provided both numerically and visually (p = 0.023). Overall, doctors had higher numerical skills than their patients (p = 0.001).
|
Applicability
Comparison of numeracy for both the patients and doctors and their inferences. Randomisation with stratification.
Limitations:
Small sample size. Statistical section was not adequate and difficult to interpret.
The data analysis was limited by the method used according to data type.
Baseline characteristics: The patients were older and less educated than the general population.
Outcome was based on inference and perception, not actual behaviour.
|
Ruiz et al., 2013, USA (28).
|
RCT at an outpatient clinic with 120 male participants
|
CVD among patients with intermediate or high cardiovascular risk.
Each participant was compensated with 30 dollars.
|
Your Cardiovascular Risk Score is a computer-based tutorial, which contains a sequential presentation of information regarding risk factors for coronary disease, their calculated absolute 10-year CVD (Framingham) and a presentation of individualised risks. The risk of a CVD is presented in three formats: frequencies, percentages or frequencies with icon arrays (red and black male stick figures). The study assessed risk understanding and knowledge by questionnaires immediately (T1), after 20 minutes (T2) and 2 weeks after the intervention (T3). T1 and T2 assessed perception of importance/seriousness, intent to adhere, and self-efficacy. T3 also concerned self-reported adherence.
The numeracy and graph literacy were also assessed.
|
Icon arrays may impair short-term recall of cardiovascular risk. Accuracy was inferior with frequencies + icon arrays compared to percentages or frequencies at T2 (p = 0.001). Patients with high graphical literacy performed better than those with low graphical literacy at all times.
|
Applicability
The patients were at high risk of CVD, which may had a positive influence on their motivation for the risk communication assessment, as they may had to make a life changing decision.
Statistical analyses explored possible effects of confounding covariates.
The person analysing the data was blinded.
Completion rate by the patients was high (88%).
Limitations:
Small sample size and from one clinic.
No sample size/power calculation.
Participants had baseline differences.
Due to risk perception being self-reported there was no measurement of actual adherence, life-style changes or medical treatments.
Short follow up period.
The use of two icon arrays: one for actual risk and one for ideal risk may have caused increased cognitive load and thereby reduced encoding.
The patients were paid to participate in the study which may change the incitement to attend.
|
Nieuwkerk et al., 2012, The Netherlands (29).
|
RCT with two outpatient clinics and 201 patients.
|
Patients with indication for statin therapy for primary or secondary prevention of CVD.
|
Extended care (EC) with nurse-led visual cardiovascular risk factor counselling compared to routine care (RC) at baseline and after 3, 9 and 18 months.
Patients in the EC group received multifactorial risk-factor counselling, and a personalised risk-factor book.
The book showed modifiable and unmodifiable individual risk factors, a graphical presentation of the calculated absolute 10-year CVD risk (Framingham). It was also showing the target risk that could be reached if all modifiable risk factors were optimally treated and the most recent ultrasound image of the patient’s carotid artery
Outcome measurements: Statin adherence, quality of life, symptoms, smoking status, blood lipids and the thickness of the carotid intima.
|
Statin adherence was higher (p < 0.01) and anxiety was lower (p < 0.01) in the EC group. LDL was lower in the EC group compared to the RC group (p = 0.024). Intima thickness decreased from baseline in both groups (p < 0.01). Multifactorial cardiovascular risk-factor counselling resulted in higher levels of adherence to lipid-lowering medication and lower LDL cholesterol concentrations in primary prevention patients, without increasing the patients’ anxiety compared to RC.
|
Applicability
Randomisation by computer to obtain equal baseline characteristics. Intention to treat analysis.
Power calculation of sample size.
Higher levels of self-reported adherence to lipid-lowering medication was significant and correlated with lower concurrent LDL cholesterol (p = 0.001), thereby supporting the validity of self-reported adherence.
Limitations
There was a difference in baseline risk perception score between the groups.
All participants had equal amounts of visits with the study nurse practitioner, but the extra time in the EC group was on average 30 minutes per visit. The positive results might have been affected by the prolonged interpersonal contact instead of being a result solely based on the risk-factor book.
|
Shukla et al., 2012, USA (30).
|
Randomised prospective study with 100 patients.
|
Cataract patients at the department of ophthalmology.
|
Patients were randomised into one of four groups: 1) Conventional verbal information; 2) conventional verbal information plus second-grade reading level brochure; 3) conventional verbal information plus eighth-grade reading level brochure; 4) conventional verbal information plus an educational DVD made for understanding cataract surgery.
All patients completed a multiple-choice questionnaire (MCQ) with 12 questions and four possible answers for each.
The MCQ revealed understanding of
surgical procedure, its benefits, its
foreseeable and unforeseeable risks, and the alternatives to cataract surgery.
|
Patients in group 2 and 4 scored higher in understanding than patients in group 1 or 3 (p < 0.001). The highest education level had no effect on scores (p > 0.05). Thus, concise informed information sheets at lower reading grade levels and videotape presentation optimised the understanding of the risks, benefits, and treatment alternatives to cataract surgery.
|
Applicability
Information by video is a reproducible and low cost procedure with potential for implementation in the primary sector.
The education level of the patients was assessed.
Limitations
No power calculation of sample size.
Uneven baseline characteristics of groups e.g. gender and education level. The MCQ was not validated.
|
McCaffery et al., 2012, Australia (31).
|
A randomised experimental study with 120 participants.
|
Adults attending government sponsored basic adult literacy and numeracy classes. They volunteered to participate in the study.
|
The target was to test optimal graphic risk communication formats when presenting small probabilities using graphics with a denominator of 1000. The experimental computer-based manipulation compared three types of graphics; bar charts and pictographs with blocks or dots across horizontal or vertical orientation. The numerator size was divided into three groups: small < 100, medium 100–499 and large 500–999. Participants were asked two questions concerning the treatment of the medical condition “X”. One focussing on gist knowledge and one on verbatim knowledge.
Three trainings were completed to ensure that the participants understood the tasks, and how to record their responses before the trial.
|
For small numerators, pictographs resulted in fewer errors than bar charts.
For medium and large numerators, bar charts were more accurate.
Accuracy on the gist task was very high across all conditions (> 95 %). Vertical formats were processed slightly faster than horizontal graphs with no difference in accuracy. Most participants preferred bar charts (64 %); however, there was no relationship with performance.
|
Applicability
Socioeconomic deprivation among the participants can be seen as a limitation as they do not represent the population or as a strength because they represent a group that is in most need for better understanding in health-related issues.
The outcome shown as probabilities is a common way of communicating risk and is relevant to the primary care sector.
Limitations:
No direct measure of the participants’ literacy or numeracy levels (average age for leaving school was 16.7 years).
Computer setting, without real patient-physician contact.
No information about the participants’ medical history.
|
Zikmund-Fisher et al., 2012, USA (32).
|
Randomised study with a quasi-factorial design and 4198 participants from a survey panel of internet users.
|
A fictive scenario about two hypothetical treatments for thyroid cancer. Tested by internet users without the disease.
|
The study evaluated eight different animated risk graphics presented by icons arrays (blocs). They were viewed on a PC screen that incorporated different combinations of three basic animations: 1) building risk one unit at a time, 2) settling scattered risk into a grouping and 3) shuffling scattered risk to reinforce randomness. Participants received all risk information in 1 out of 10 possible pictograph formats.
Outcome: To test if animated icon array pictographs, displaying risks of side effects, could improve participants’ ability to select the treatment with the lowest risk profile, as compared with seeing static images of the same risks.
Outcome measurements: The ability of the participants to choose the less risky treatment (choice accuracy), gist knowledge of side effects (knowledge accuracy), and graph evaluation ratings, controlled for subjective numeracy, and need for cognition.
|
No animations significantly improved any outcomes, compared to static grouped icon arrays. The most animations showed significant performance degradations (p < 0.02). Displays with scattered icons (static or animated) performed particularly poor unless they included a settled animation that allowed users to see event icons grouped. Static pictographs that grouped event icons at the bottom of the array consistently resulted in an optimal treatment choice, higher knowledge accuracy and better graph evaluation ratings. The most complex graphics were least preferred by the participants.
|
Applicability
Large sample size, with diversity and without specific diseases. The email invitations were regulated to ensure stratification for sub samples. Calculation of the participants’ subjective numeracy.
Limitations:
The internet panel was given a hypothetical medical treatment scenario. This may have had an influence on the participants’ motivation to engage in the task.
Only representative for a population of internet users.
No demographic data on drop outs and non- responders.
Completion rate 67.7 %.
Graphs of the two conditions were presented side-by-side, making it possible that the dual animation affected the outcome. The animated blocks finished appearing in one of the two arrays before the other one, creating a longer motion cue, which may have affected treatment choice.
|
Hawley et al., 2008, USA (3).
|
Randomised trial with 2412 participants drawn from a survey panel of internet users.
|
An online hypothetical medical decision-making scenario about CVD.
Setting: General practice.
|
Imaginary scenario in general practice with a choice between two different types of medication to avoid a bypass surgery. One treatment was designed as superior according to its risk profile and beneficial effects. Numerical risk information was given in one out of the following six graph formats; bar graph, pictograph, modified pictograph (sparkplug), pie chart, modified pie graph (clock graph) or in a table.
The aim was to evaluate what impact these six graphical formats had on answers about treatment risks and benefits.
The outcome measurements: Verbatim knowledge (the ability to correctly read numbers from graphs) and gist knowledge (the ability to identify the essential points of the information presented).
|
All formats were positively received, and pictographs were trusted by respondents with both high and low numeracy. High Verbatim and gist knowledge where associated with making a medically superior treatment choice (p < 0.01).
Viewing a pictograph was associated with both adequate verbatim and gist knowledge, especially for individuals with lower numeracy.
Respondents with higher numeracy answered more of the questions correctly for both verbatim and gist knowledge regardless of graph type (p < 0.007), compared to respondents with low numeracy, none of the graph types were associated with making a correct treatment choice.
Pictographs were the best format for communicating probabilistic information, particularly among individuals with lower numeracy.
|
Applicability
Email invitations were adjusted to ensure stratification for sub samples. Large sample size.
Numerical understanding was translated into the understanding of consequence.
Limitations:
Participants were not personally affected by the risk presented, this may have affected the way risk was interpreted.
Data was only representative for people able to use the internet.
Dropout (23.5 %), may be due to participants who did not understand the graphs. This has not been explored further.
No real patient-physician contact regarding delivery of medical information.
Numeracy was evaluated by using a validated method, however the questions used for gist and verbatim knowledge were not confirmed validated.
|
Goodyear-Smith et al., 2008, New Zealand (33).
|
Question-naire and telephone interviews with 188 patients invited through
four family practices.
|
Patients with a pre-existing heart disease and users of statin.
|
Patients were interviewed about their preference for methods expressing the preventive benefit of a hypothetical medication. Benefits were expressed in numerical formats (relative risk, absolute risk, number needed to treat, odds ratio and natural frequency) and one graphical (bar chart).
The outcome measurements: Could information presented in a different way encourage the patient to take the medication daily, which method was preferred to express the benefit of the medication and if the patient preferred positively or negatively framed information.
|
No matter how the risk was expressed, most of the patients (67-89 %) indicated that they would be encouraged to take the medication. A large group of the patients (68 %) preferred one method of expressing benefits over the others, but 32 % of the patients could not decide which presentation they preferred. More than half (57 %) preferred the information presented graphically (p < 0.001). The second most preferred option (19%) was relative risk. Most patients (90 %) preferred positive framing (description of the benefits of treatment) above negative framing (description of the harm of not being treated).A graphical representation of the benefits was the method patients preferred the most.
|
Applicability
Study sample included patients with a known disease, who are familiar with taking medication every day, thus consider positive effects or side effects of medication daily.
Limitations:
Not randomised or controlled design.
Small study sample without power calculation.
The interviewer was not blinded to the type of format the patient was evaluating.
Low response rate: 53 % (100 patients).
The questionnaire was not validated.
The study was not done at the point of true decision making.
The study presumed that the preference for a given format of explanation reflected the ease of the patient to understand the information presented.
|