This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Most displays of laboratory test results include a standard reference range. For some patients (eg, those with chronic conditions), however, getting a result within the standard range may be unachievable, inappropriate, or even harmful.
The objective of our study was to test the impact of including clinically appropriate goal ranges outside the standard range in the visual displays of laboratory test results.
Participants (N=6776) from a demographically diverse Web-based panel viewed hypothetical glycated hemoglobin (HbA1c) test results (HbA1c either 6.2% or 8.2%) as part of a type 2 diabetes management scenario. Test result visual displays included either a standard range (4.5%-5.7%) only, a goal range (6.5%-7.5%) added to the standard range, or the goal range only. The results were displayed in 1 of the following 3 display formats: (1) a table; (2) a simple, two-colored number line (simple line); or (3) a number line with diagnostic categories indicated via colored blocks (block line). Primary outcome measures were comprehension of and negative reactions to test results.
While goal range information did not influence the understanding of HbA1c=8.2% results, the goal range only display produced higher levels of comprehension and decreased negative reactions to HbA1c=6.2% test results compared with the no goal range and goal range added conditions. Goal range information was less helpful in the block line condition versus the other formats.
Replacing the standard range with a clinically appropriate goal range could help patients better understand how their test results relate to their personal targets.
In an effort to facilitate greater patient involvement in the management of their health, hospitals and health care systems have increasingly provided patients with access to their electronic health records (EHRs) [
Individuals who manage chronic conditions face an additional barrier to understanding and effectively using their test results: inappropriate reference ranges. The standard range commonly presented as part of test result communications represents the distribution of values commonly observed in a healthy population [
In addition to the use of clinically appropriate goal ranges, use of visual displays could help increase patient sensitivity to variations among out-of-range results. In a previous study by our research group, we tested the impact of presenting laboratory test results via 3 number line formats versus a standard table format on participants’ sense of urgency and desire to contact their health provider [
These issues raise the question of how can test results be communicated to patients in ways that help them better understand how their result compares to the target range most relevant to their self-management and treatment decision making. To the best of our knowledge, there has been no research examining whether and how individual- or disease-specific goal range information should be incorporated into the returned laboratory test results for patients such as these. Inclusion or exclusion of different combinations of these reference standards might improve patients’ comprehension of the test value and reduce unnecessary negative reactions, such as discouragement or urgency to contact their health care provider when urgency is unnecessary.
We conducted a Web-based experiment in which respondents imagined receiving HbA1c test results through an EHR patient portal as part of the ongoing management of their type 2 diabetes. This study was designed to answer four key questions:
Does the inclusion of goal range information improve comprehension of the test results?
Does the inclusion of goal range information reduce unnecessary negative reactions to test results that are outside of the standard range, but near their goal range?
Is it better to include the goal range information in addition to, or in place of, the standard range?
Does the display format (eg, table vs visual number line) change the impact (if any) of including goal range information in the test result display?
Utilizing the principle “less is more,” which has been shown to apply in health communication [
Data were collected through Qualtrics survey software (Qualtrics; Provo, UT) from a nationwide sample of US adults through Survey Sampling International (SSI). Participants were recruited over a 2-month period from August to October 2015.
Participant eligibility was determined through SSI using a probability-weighted random process based on sample requirements. We established quotas on respondent age (33% aged 21-39 years, 33% aged 40-49 years, and 33% aged ≥60 years), gender (50% females), and race or ethnicity (14% African American, 14% Hispanic, and 4% Asian American people) to approximate the distribution of these characteristics in the US population. However, we oversampled individuals with diabetes to ensure that we did not have an overly healthy sample and to evaluate whether experience managing diabetes moderated the impact of the goal presentation format. SSI participants were routed to the survey via the sampling algorithm until all quotas were achieved.
Participants were asked to imagine that they had recently visited their doctor’s office to discuss the management of their type 2 diabetes, during which their doctor had highlighted that people with type 2 diabetes should try to have HbA1c values within a target or goal range of 6.5%-7.5%. Participants were then asked to imagine that in the intervening 3 months, they did their best to follow their doctor’s recommendations (eg, exercising regularly and eating healthy). Then, 3 months prior to their next appointment, the patients underwent some blood tests and viewed the results of these tests a day later via a Web-based EHR portal.
We tested 3 between-subjects factors (varied independently) to examine the impact of including goal range information across different presentation formats on patient reactions to their test results.
The second factor was HbA1c test value. Participants were randomly presented with an HbA1c test value of 6.2%, which fell between the standard range and the goal range, or a value of 8.2%, which was higher than both the standard and goal ranges. The third factor was the display format
We included 2 measures to assess how well participants understood their test result in relation to their goal range. For the relative location measure, we asked, “Where was your test result compared to your goal range?” with “higher than the goal range,” “within the goal range,” “below the goal range,” and “I don’t know” as response options. For the future location measure, we asked, “At your next test, what do you think your next test result should be, as compared to this test result?” using a 9-point Likert scale response option with “A lot lower” and “A lot higher” as the anchor labels and “About the same” as the midpoint label. “I don’t know” was also included as an additional response option.
We included 2 measures to assess participant reactions to their test result: one measuring how discouraged they would be by their test result and one assessing whether and when they would contact their doctor about their test result. For the discouraged measure, we asked, “How discouraged or encouraged do you feel about this test result?” using a 6-point Likert scale response option with “Very discouraged” and “Very encouraged” as the anchor labels with an additional “I don’t know” response option. For the urgency measure, we asked, “How soon do you need to speak to your doctor regarding these results?” with “Immediately,” “Within a few weeks,” “At your next appointment in 3 months,” and “I don’t need to speak to my doctor about these results” as response options.
We asked participants about their age, gender, race and ethnicity, education, and whether they have diabetes, and if so, what type.
All data were collected anonymously so that the researchers had no way to learn the identity of the participants. A unique identification number provided by SSI was contained in the redirected URL, which identified participants and prevented them from completing the study multiple times. This study was deemed exempt by the University of Michigan Health Sciences and Behavioral Sciences Institutional Review Board.
Goal presentation and display formats for 6.2% glycated hemoglobin (hemoglobin A1c) test value; labels indicate the display format and goal presentation.
Responses to the relative location measure were recoded as “1” to indicate a correct response if participants responded “below the goal range” in the 6.2% HbA1c test result condition or “higher than the goal range” in the 8.2% HbA1c test result condition. All other responses were recoded as “0” to indicate a failure to know where their test value was in relation to the goal range. To assess whether participants had the gist of where their next test value should be, future location responses were recoded as “1” if they were above the midpoint of the scale in the 6.2% HbA1c test result condition and below the midpoint in the 8.2% condition. All other responses were recoded as “0.” The results are substantially the same, if not stronger (ie, larger effect sizes), if “about the same” is coded as “1”. The one exception is that having diabetes is associated with an increased comprehension of the future location for the goal presentation and display format logistic regression analysis. Responses to the discouraged and urgency measures were reverse coded, such that higher scores indicated greater discouragement and urgency, respectively. We recoded gender (0=male, 1=female), race (0=white, 1=nonwhite), and diabetes status (0=no diabetes, 1=diabetes).
We report percentages for the relative and future location measures and descriptive measures for the discouraged and urgency measures across the different factors. We used chi-square analyses to test for differences in percentages and independent sample
Of all the participants who initiated the study, 83.09% (6781/8161) completed it. In addition, 14 responses were dropped due to a reported age <18 years old, and 1 response was dropped due to a reported age of 586.
In univariate analyses of participants receiving HbA1c=8.2% in table form, neither goal presentation nor display factors significantly affected any of the outcomes (all
Among participants who received tabular displays of HbA1c=6.2% results (which fell between the standard and goal ranges), however, goal presentation format had a significant impact on comprehension. As shown in
The logistic regression analyses of participants receiving test results in table format (
Sample characteristics (N=6766).
Characteristic | Valuea | |
Age, mean (SD) | 49.1 (15.8) | |
Male | 3299 (48.88) | |
Female | 3435 (50.90) | |
Transgender or other | 15 (0.22) | |
Hispanic (any race) | 892 (13.26) | |
White | 5294 (78.24) | |
African American | 1002 (14.81) | |
All other | 654 (9.67) | |
<High school | 135 (2.00) | |
High school only | 1065 (15.78) | |
Some college or trade | 2458 (36.41) | |
Bachelor’s degree | 2005 (29.70) | |
>Bachelor’s degree | 1087 (16.10) | |
No diabetes | 3620 (53.79) | |
Type 1 diabetes | 497 (7.38) | |
Type 2 diabetes | 2613 (38.83) | |
Standard range only | 2253 (33.30) | |
Goal range added | 2219 (32.80) | |
Goal range only | 2294 (33.90) | |
6.2% | 3390 (50.10) | |
8.2% | 3376 (49.90) | |
Table | 2251 (33.27) | |
Simple line | 2224 (32.87) | |
Block line | 2291 (33.86) |
aResults reported only for those respondents who completed each question or measure.
bRespondents could mark more than one race.
Effect of providing goal range information in table format, by goal presentation type and glycated hemoglobin (HbA1c) test result; asterisks indicate statistically significant differences between the 2 bars. Std range: standard range.
Logistic regression and ordered logistic regression results showing predictors of outcome measures, table condition only.
Predictors | Relative location | Future location | Discouraged | Urgency | ||||||
ORa (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |||||||
Standard range only | Reference | Reference | Reference | Reference | ||||||
Goal range added | 4.98 (3.46-7.17) | <.001 | 1.69 (1.23-2.32) | <.001 | 0.51 (0.39-0.67) | <.001 | 0.77 (0.59-1.00) | .049 | ||
Goal range only | 6.83 (4.77-9.76) | <.001 | 2.52 (1.86-3.42) | <.001 | 0.47 (0.35-0.59) | <.001 | 0.74 (0.57-0.95) | .02 | ||
6.2% | Reference | Reference | Reference | Reference | ||||||
8.2% | 12.07 (8.30-17.54) | <.001 | 4.30 (3.13-5.91) | <.001 | 4.13 (3.12-4.46) | <.001 | 2.16 (1.63-2.86) | <.001 | ||
Goal range added × 8.2% | 0.25 (0.15-0.40) | <.001 | 0.57 (0.37-0.87) | .01 | 1.40 (0.95-2.05) | .09 | 1.24 (0.84-1.84) | .28 | ||
Goal range only × 8.2% | 0.15 (0.09-0.24) | <.001 | 0.49 (0.32-0.75) | <.001 | 1.67 (1.15-2.44) | .01 | 1.16 (0.79-1.71) | .46 | ||
Diabetesb | 0.64 (0.52-0.77) | <.001 | 1.03 (0.87-1.23) | .72 | 0.52 (0.44-0.61) | <.001 | 0.98 (0.84-1.15) | .83 | ||
Agec | 1.02 (1.01-1.03) | <.001 | 1.00 (0.99-1.00) | .54 | 1.01 (1.01-1.02) | <.001 | 0.98 (0.97-0.99) | <.001 | ||
Genderd (female) | 1.74 (1.44-2.11) | <.001 | 1.03 (1.03-1.47) | .02 | 1.16 (0.99-1.35) | .06 | 0.84 (0.72-0.99) | .04 | ||
Racee | 0.70 (0.56-0.87) | .002 | 1.00 (0.81-1.24) | .99 | 1.01 (0.84-1.22) | .92 | 1.08 (0.90-1.32) | .39 | ||
Educationc | 1.11 (1.05-1.17) | <.001 | 1.03 (0.98-1.08) | .24 | 0.99 (0.94-1.03) | .57 | 1.08 (1.04-1.13) | <.001 | ||
Constant | 0.03 (0.02-0.05) | <.001 | 0.28 (0.16-0.47) | <.001 | N/Af | N/A | N/A | N/A |
aOR: odds ratio.
bDiabetes (0=no, 1=yes).
cAge and education treated as continuous variables.
dGender (0=male, 1=female).
eRace (0=white, 1=nonwhite).
fN/A: not applicable.
Given that providing goal information to participants receiving test results via tables only influenced outcomes among those viewing HbA1c=6.2% results, we focused only on these conditions when comparing optimal formats (ie, table vs simple line vs blocks line) for presenting goal information. As shown in
The logistic regression analyses (
Effects of presenting goal information to patients viewing glycated hemoglobin 6.2% test result: goal presentation and display format; asterisks indicate statistically significant differences between the 2 bars. Std Range: standard range.
Logistic regression results showing goal presentation, presentation format, and demographics as predictors of outcome measures, 6.2% glycated hemoglobin test value condition only.
Predictors | Relative location | Future location | Discouraged | Urgency | |||||||||||
ORa (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | ||||||||||||
Standard range only | Reference | Reference | Reference | Reference | |||||||||||
Goal range added | 4.85 (3.38-6.96) | <.001 | 1.73 (1.25-2.39) | <.001 | 0.45 (0.34-0.59) | <.001 | 0.74 (0.56-0.96) | .03 | |||||||
Goal range only | 6.6 (4.65-9.44) | <.001 | 2.65 (1.94-3.63) | <.001 | 0.39 (0.30-0.51) | <.001 | 0.72 (0.55-0.93) | .01 | |||||||
Table | Reference | Reference | Reference | Reference | |||||||||||
Simple line | 1.02 (0.68-1.54) | .92 | 1.01 (0.73-1.42) | .93 | 0.85 (0.65-1.11) | .23 | 0.92 (0.71-.20) | .55 | |||||||
Block line | 0.61 (0.39-0.97) | .04 | 0.60 (0.41-0.86) | .01 | 1.17 (0.90-1.53) | .25 | 0.83 (0.64-1.08) | .16 | |||||||
Goal range added × simple line | 1.17 (0.70-1.95) | .54 | 1.17 (0.73-1.85) | .51 | 1.16 (0.78-1.70) | .46 | 0.99 (0.68-1.46) | .97 | |||||||
Goal range added × block line | 1.07 (0.62-1.86) | .82 | 1.02 (0.63-1.67) | .93 | 1.56 (1.07-2.28) | .02 | 1.22 (0.83-1.78) | .31 | |||||||
Goal range only × simple line | 0.99 (0.60-1.64) | .97 | 1.19 (0.76-1.86) | .45 | 1.36 (0.94-1.97) | .11 | 1.20 (0.83-1.74) | .34 | |||||||
Goal range only × block line | 1.74 (1.01-2.99) | .045 | 2.32 (1.45-3.71) | <.001 | 1.29 (0.89-1.86) | .18 | 1.09 (0.76-1.58) | .64 | |||||||
Diabetesb | 0.64 (0.55-0.75) | <.001 | 0.77 (0.66-0.89) | <.001 | 0.31 (0.27-0.35) | <.001 | 0.74 (0.65-0.84) | <.001 | |||||||
Agec | 1.01 (1.01-1.02) | <.001 | 0.97 (0.96-0.97) | <.001 | 1.00 (0.99-1.00) | .049 | 0.97 (0.97-0.98) | <.001 | |||||||
Female genderd | 1.56 (1.33-1.82) | <.001 | 0.90 (0.77-1.05) | .19 | 1.04 (0.91-1.18) | .57 | 0.79 (0.69-0.89) | <.001 | |||||||
Racee | 0.76 (0.63-0.92) | .01 | 1.02 (0.85-1.22) | .85 | 0.96 (0.82-1.12) | .62 | 1.20 (1.03-1.40) | .02 | |||||||
Educationc | 1.11 (1.06-1.16) | <.001 | 1.04 (1.00-1.09) | .07 | 0.98 (0.95-1.02) | .29 | 1.03 (0.99-1.06) | .14 | |||||||
Constant | 0.04 (0.03-0.07) | <.001 | 1.38 (0.87-2.19) | .17 | N/Af | N/A | N/A | N/A |
aOR: odds ratio.
bDiabetes (0=no, 1=yes).
cAge and education treated as continuous variables.
dGender (0=male, 1=female)
eRace (0=white, 1=nonwhite)
fN/A: not applicable.
Demographic covariates remained significant predictors across the 4 outcome measures (see
The regression results presented in
Our data suggest that providing people with test results displays (tabular or visual) that include goal range information can alter their perceptions of their test results in important ways. While perceptions were generally unaffected by format when the result was above both the standard and goal ranges, perceptions were sensitive to format when the result was above the standard range but below the goal range. Comprehension of the below-target nature of this result was higher when goal information was explicitly included in participants’ test result tables or visual displays. Furthermore, inclusion of goal information in the display reduced perceived discouragement about the presented results.
Our data also show that removing the standard range and substituting it with a single goal reference range seems superior to simply adding goal range information along with the standard range values. Comprehension was highest and discouragement and urgency were lowest when the goal range information was presented in lieu of the standard range information. This suggests that it is difficult for people to put aside information about the standard range—which is normed based on the total, mostly healthy, population—even when more personalized goal information is easily available. As a result, the inclusion of these standard reference points (which are less relevant in this particular situation) may undermine patients’ ability to manage their chronic conditions and may expose them to harm when aggressively trying to achieve test results within the standard range [
Fundamental principles of both visual design and information evaluability suggest that the dominance of the goal only substitution condition is due to the fact that the inclusion of more than 1 reference range produces confusion about which comparator is most relevant to understanding where the patient’s test value should be [
One limitation of our study is the use of a hypothetical scenario. While participants did not receive actual test results, approximately half of our sample had the medical condition described in the scenario (diabetes) and would likely have experience receiving HbA1c test results. While we found the same pattern of results for participants with and without diabetes, participants with diabetes who received HbA1c=6.2% results were less likely to report that their values were too low, but these participants also exhibited decreased discouragement and urgency. One possible explanation for this finding is that their experience with repeatedly being told that their HbA1c goal should be below 7.0% has led them to adopt the standard range as the norm that they should be striving to attain, even when an alternative goal range has been provided. Another possibility is that participants with diabetes were relying on their real-life goal ranges, which may have been different from the one provided in the scenario, or that they recognize that not all persons with type 2 diabetes will experience adverse outcomes with an HbA1c of 6.2%. This explanation may account for the overall smaller percentage of participants with diabetes who were discouraged about their test result or felt a need to contact their health care provider immediately.
As more and more patients receive their test results via Web-based patient portals, it is becoming increasingly important that patients should be able to find their results meaningful and that we do not cause unnecessary distress or discouragement to patients. Current approaches to presenting laboratory test results to patients appear to be particularly problematic for many patients, such as those with chronic conditions, who may have personal target goals that differ from those relevant to healthy adults. For these patients, the standard range commonly shown is not necessarily where we want patient results to be. Providing goal range information in place of the standard range may be one step toward reducing these problems with EHR systems; however, challenging discussions would need to occur regarding the pros and cons of who should determine the goal range information (ie, health systems, EHR or portal vendors, expert panels, individual physicians, and/or patients) or what the goal ranges should represent (eg, broader goals for people with a chronic condition vs individualized goals). More research is needed to determine additional features that may further improve the interpretability of laboratory test results.
Logistic regression and ordered logistic regression results showing goal presentation, diabetes status, and demographics as predictors of outcome measures, 6.2% A1c test value condition only.
analysis of variance
electronic health record
glycated hemoglobin
Survey Sampling International
This work was previously presented at the Annual Meeting of the Society for Medical Decision Making, Vancouver, BC, Canada; October 24, 2016. Funding for this research was provided by a grant from the US Agency for Healthcare Research and Quality to BJZF (R01 HS021681). The funding agreement assured the authors’ independence in designing the study; in the collection, analysis, and reporting of the data; and in the decision to submit the article for publication. The authors acknowledge the assistance of Sandeep Vijan, MD, Kenneth Langa, MD, and Beth Tarini, MD, in determining the appropriate display ranges and categorization schemas for each of the types of test results used in this study. In addition, we acknowledge the graphic design efforts of Grace Bienek in creating the images used in this study. Lastly, we are grateful for the guidance and inspiration provided by the patient members of our research team: Margaret Newton, Stephanie Burke, and James Piazza.
None declared.