Table 1.

Features of the presentation of information on health care report cards from previous studies.

Category	Recommendations and results
Evaluative table with symbols	Consider using a table design such as the “evaluative table with stars” rather than a bar chart [29]
	Evaluative tables using words or stars are superior to numerical tables [29]
	Physicians preferred formats that used traffic light symbols to code the value of indicators (numerical table with traffic lights) [30]
Tables without symbols	Graphic displays were more helpful to users than text-only tables [31]
Bar charts	Bar charts were commonly used (43% of public reporting websites) [32]
Bar charts without symbols	Comprehension was lowest when data were presented in bar charts [30]
	Standard bar charts were not well-liked by respondents and led to the lowest levels of comprehension [29]
Bar charts with symbols	Symbols and bar charts should be used [31]
	A combination of bar charts and star ratings facilitated correct interpretation by users [32]
	Adding stars to bar charts increases comprehension significantly [33]
Symbols	Participants liked to use symbols to identify the best surgeon [31]
	Physicians preferred formats that used symbols (eg, traffic lights) [30]
	Star-only formats should be used in preference to numerical values [34]
	Only important information should be made easier to evaluate using symbols [35]
Evaluative word labels	Adding evaluative labels to bar charts did not increase comprehension [33]
Highlighting	Color-coding important information improves comprehension [36]
	Highlighting information about quality resulted in greater understanding [37]
	Presentation formats which highlighted key messages increased comprehension [38]
Order of providers	Physicians prefer presentation formats that combine individual indicator values with evaluative features such as rankings [30]
	Comprehension of respondents who were low in numeracy was significantly improved by the ordered compared to the unordered condition [35]
	Providers should be ranked by performance [12]
	Ranking plans by performance significantly decreased errors in interpreting data [33]
	Ranking by performance increased the frequency with which users chose higher-performing services [15]
	Providers should be ranked in descending order of quality, as this was valued by participants and increased their comprehension [36]
	One of the more powerful display strategies is to rank providers in terms of performance [33]
	When providers were ordered alphabetically participants were more likely to make effective use of the data (ie, choose the best provider) than when providers were ordered by performance [32]
High values indicate good performance	Performance data should be displayed such that high values always represent high performance [35]
	Numeric tables and bar charts often led respondents to conclude that the worst performing nursing homes (those with the higher percentages) were the best, notwithstanding the warning label at the top [29]
State explicitly whether high or low values indicate good performance	It should be stated explicitly whether high or low values indicate good performance, regardless of the direction of the scale [8,36]
Incomplete data (“N/A” as a value)	Incomplete data (missing values) have a negative influence on provider assessment and the potential to influence a decision [29]