Not all development projects succeed. This paper describes and evaluates a failed project.

a well-designed document

For a document to be acceptable to people — to be picked up, read and used rather than picked up, glanced at, and thrown in the bin — it must meet certain criteria.

Ideally, to avoid the bin, it must:

be credible

respect its readers

be attractive

be physically appropriate for its readers

be socially appropriate for its readers.

Ideally, to be read and used, it must:

be easy to use

be efficient to use

lead to a productive outcome from the reader's point of view (Sless 2004).

The document design project reported here has consistently failed to meet these criteria. Not surprisingly, many of its readers find it unacceptable and don't use it. The document in question is Consumer Medicine Information (CMI).

Fully illustrated, with all the available data collected during this project, this is the most comprehensive CMI case history in the field.

Note: This paper was originally researched and written by Cheryl Edridge of CRIA in 2003. It was updated by David Sless and Ruth Shrensky in May 2009

List of abbreviations

APAC — Australian Pharmaceutical Advisory Council
CHF — Consumers' Health Forum
CMI — Consumer Medicine Information
CRIA — Communication Research Institute of Australia (CRI since 2006)
EDWG —Electronic Delivery Working Group
MIC — Medicines Information for Consumers
QARG — Quality Assurance Reference Group
PHARM — Pharmaceutical Health and Rational Use of Medicines
PGA —Pharmacy Guild of Australia
PSA — Pharmaceutical Society of Australia

Acknowledgements

APACWe would like to acknowledge the contributions of the following people and organisations to this project and thank those who tested the CMI.

Deborah Monk, Heather Friedel, Janne Graham, Jenny Chu, Judith Tasker, Juliet Seifert, Kim Bessell, Lloyd Sansom, Marlene Arens, Rosemary Knight, Susan Parker, Sylvia Roins.

APAC, AstraZeneca, CHF, Commonwealth Department of Health & Aged Care, EDWG, Merck, Novartis, Pfizer, QARG, Sigma.

The purpose of CMI

CMI is designed to provide accurate, up-to-date information to consumers about prescription medicines and pharmacy-only medicines. Individual CMI leaflets (CMIs) should provide all the information users might need about a specific medicine product. CMI is provided in the wider context of health information as a communication and counselling link between healthcare providers and medicine consumers. The content of CMI is taken from the Product Information (PI) of the medicine written by the medicine's sponsor or manufacturer; nothing may appear in the CMI that is not in the PI, so consumers can be confident that they are being given reliable information. Also, nothing may appear in the CMI that can be seen as an advertisement or endorsement of the brand.

Individual CMIs are developed by pharmaceutical companies. Ideally, whenever a prescription medicine is dispensed at a community pharmacy*, the pharmacist should give the relevant CMI to the medicine user, pointing out important details of usage.

(*A 'community pharmacy' is a pharmacy serving the public—a high street chemist as opposed to a hospital or specialised pharmacy. All references to pharmacies in this paper refer to community pharmacies.)

Mass customisation of documents

The system used to generate pharmacy CMI is an example of a system for mass customisation of documents. Such systems are widely used, and in many fields they are superceding the mass production technology which has been used to print documents for the last 500 years. In the older technology, a single document is reproduced many times over. In the newer technology a set of rules is used to drive the printer so that every single document is customised for the eventual user. A simple example of this mass customisation is the form letter which is individually addressed to a particular recipient, though the content remains the same for all recipients. A more sophisticated version is the bill, where not only the name and address of the recipient but the amounts and descriptions of the items are customised. Many examples exist of even more sophisticated customisation involving complex logical rules that make decisions about specific content on individual letters, depending on multiple characteristics of the recipient. Two examples are shown below, Figures 1 and 2.

Figures 1 & 2. click to enlarge.

These designs were developed using CRI's professional information design process, shown in Figure 3.

Figure 3.

The design of CMI for pharmacies

The appearance of CMI in general use today is highly constrained by the document customisation system used for printing in pharmacies at the point of dispensing. The system is built into the pharmacists' dispensing software, which has very limited print control features. In some cases, pharmacists are still using dot matrix printers. At best, the software can output the crude word-processing code designed for laser printers in the mid-1980s, using the small number of resident fonts in pharmacists' printers. The fonts used for CMI have to be chosen from those most commonly available on dot matrix printers or primitive laser printers.

In the case of dot matrix design, the choice of font is restricted to two sizes of the same dot structure, available in regular, bold, and italics. Line spacing is single or double. There is no control over pagination or columns.

In the case of laser printing design, the fonts are restricted to Times and Helvetica, these being the most commonly available fonts on laser printers. These printers have slightly more flexibility in font sizes, line and paragraph spacing, but not much. Letter spacing is governed by simple proportional spacing rules. Altogether, the system has fairly crude specifications, not unlike Microsoft Word 3.0.

However, anticipating the possibility that CMI might one day be implemented by professional document production systems, CMIs were designed from the start (1994) with advanced document-structuring features, using Styles and sgml/xml-like definitions. This advanced structuring has been largely ignored by CMI writers. Sadly, to this day, the crude designs which had to be developed to work on the earlier technology are still in use, and in most instances fail to take advantage of design features such as Styles which are now readily available on word processors. Indeed, implementation in most circumstances fails to meet even the crude standards set in 1994. See Figures 4 &5.

Figures 4 & 5. Click to enlarge.

The full project to develop these designs is reported as one of our case histories (Penman et al 1996). The table below summarises where that project had got to in 1996, in terms of meeting the criteria for good document design—and in 2009, it is still there.

bin-avoidance criteria	pharmacy cmi	why?
1. credible	don't know	not tested
2. respectful	don't know	not tested
3. attractive	no	technology limited
4. physically appropriate	no	technology limited
5. socially appropriate	don't know	technology limited
criteria for reading and using
6. usable	yes	tested
7. efficient	no	technology limited
8. productive	don't know	not tested

The failure to meet the first five criteria means that pharmacy CMIs, even if widely available, are unlikely to be read by most pharmacists and consumers. Nonetheless, testing of these crude designs showed that they were usable, and the target performance was set at a level that research had shown was realistic for pharmaceutical companies to reach. The resultant guidelines specified the level as follows:

• at least 90% of literate consumers should be able to find information on the CMI quickly and easily
• at least 90% of those who find the information should be able to act appropriately on the information
• thus at least 81% of literate consumers should be able to use the CMI appropriately.

These levels have remained the same through all editions of the Guidelines (Sless & Shrensky 2006).

CMI outside the pharmacy system

Beyond the procrustean constraints of pharmacy-generated CMI, some manufacturers decided to aim for a higher standard of CMI, using such things as stand-alone booklets. Free of the constraints, these CMI were able to meet all the criteria for good document design. Two examples are shown in Figures 6 & 7.

Figures 6 & 7. Click to enlarge.

These examples are among many that achieve a higher performance than pharmacy CMI, demonstrating that higher standards of CMI are possible outside the pharmacy.

Back to pharmacy CMI

Since 1994, the pharmaceutical industry, consumer groups, and other interested stakeholders have been concerned that pharmacy CMI are neither widely used nor achieve their objective of providing consumers with better information about the medicines they are taking. In particular, they are concerned that pharmacists do not routinely provide CMIs with prescription medicines and use the opportunity for counselling consumers. Certainly, pharmacists are not specifically required to supply CMIs under any specific legislation or regulation; however, they do have a professional and legal obligation to ensure patients have the required information to enable them to make informed decisions about their medicines, and CMIs are now the recommended primary method to meet these obligations (PSA 2007).

Pharmacists argue that, along with other difficulties [a companion paper to this, CMI and the pharmacists Shrensky & Edridge 2009, gives details of the pharmacists' arguments], a major barrier is the length of individual CMIs. Most CMIs comprise two pages (three to four sides), but pharmacists have been demanding one-page CMIs from the outset. This can be seen as an instance of CMI not reaching the criteria of physical or social appropriateness for pharmacists, as one group of CMI users.

It has also been clear to other stakeholders that there are credibility problems. CMI do not look authoritative, and they are not attractive: two factors known to inhibit the uptake of documents by potential readers.

One-page CMI: phase 1

The pharmaceutical industry supported the idea of shorter, more attractive, more credible CMIs as a way of increasing acceptance by pharmacists, but was reluctant to spend time and money researching and developing different CMI formats where there was little certainty that pharmacists would deliver them or that consumers would accept them.

However, in 1999, the pharmaceutical company Sigma invited CRIA to design one-page (two-sided) CMIs in pdf format for the top twenty medicines in use at the time, to demonstrate that it could be done.

Freed from the graphic constraints inherent in pharmacy CMI, CRIA opted for professional-level graphic design technology (Adobe InDesign), with the final output to be effected as pdf at the point of dispensing.

The medicines chosen and the number of words on their CMIs were as follows:

Medicine name	CMI word count
Fluvax injection	1244
Zantac	1331
Panadeine Forte	1536
Ventolin Nebules	1626
Pulmicort	1675
Ventolin Inhaler	1698
Norvasc	1731
Augmentin Forte Syrup	1789
Temaze Triphasil Tablets	1921
Atrovent Unit Dose	2019
Losec Capsules	2044
Valium	2061
Tenormin	2065
Imdur tablets	2163
Zoloft	2193
Lipitor Tablets	2206
Tenopt	2289
Coumadin Tablets	2461
Renitec	2518

The number of words was important because the largest gave the upper limit the design had to accommodate if it was to work for all the chosen CMI.

CRIA separated the information on CMI into two categories: general information that was applicable to all medicines, and information that was specific to a particular medicine. The general information was printed on one side of the sheet (see Figure 8), the specific on the other. One advantage of this was that the general information could be preprinted, so pharmacists would only have to print out the second page. The major headings were preprinted on the reverse, ready to take the specific content (see Figure 9).

Figures 8 & 9. Click to enlarge.

From previous research (Penman et al 1996) it was clear that users of a one-page CMI would find it difficult to locate information. Consequently, the new design significantly enhanced the main headings, to make it easier to navigate the document.

The layout was in landscape format, with the information given in six columns, each column dedicated to a major CMI heading so that the structure of the document could be seen at a glance. The font family chosen for the CMI was Poppl Laudatio, which is easily readable in smaller sizes and was judged to lend credibility to the content. CRIA also implemented a potential CMI logo (the Commonwealth Government had commissioned CRIA to design one). And overall, the document was designed to be attractive.

Counselling points

Observations by CRIA of pharmacists using CMI, and data collected at a CMI workshop (1998) revealed quite clearly that pharmacists and other health professionals did not know how to incorporate CMI into their counselling of consumers. To assist this process CRIA developed the idea of counselling points—small black squares which could be placed in the margin at points on the CMI where health professionals could draw consumers' attention to specific points of importance, thus providing a more focused and directed approach to counselling activity.

Figures 10 to 15 show how the content of CMI with different word lengths was fitted into the design, and what the counselling points looked like. (Note that the word counts given in the table above include the general information, 322 words, whereas the captions to the Figures below show only the number of words for the specific information page.)

Figures 10 to 15. Click to enlarge.

The new design was not tested, as Sigma had funded the project for demonstration purposes only. However, as the table below shows, attempts were made to satisfy many of the criteria for good document design.

bin-avoidance criteria	Sigma cmi	why?
1. credible	possibly	used appropriate font and CMI logo
2. respectful	don't know	not tested, but making a document more attractive can help
3. attractive	possibly	refined typographical hierarchy using InDesign and two colour printing
4. physically appropriate	possibly	reducing the size down to one leaf met the pharmacists requirements
5. socially appropriate	possibly	addition of counselling points
criteria for reading and using
6. usable	possibly better	document heading structure more obvious
7. efficient	possibly	heading structure might make search quicker
8. productive	don't know	not tested

The Pharmacy Guild was delighted with the design but was disinclined to finance further research and testing. The pharmaceutical industry was not delighted as this would mean they would have to edit or rewrite all their existing CMI. Thus the original multi-page CMIs remained in current use—and pharmacists continued to complain about the length.

In 2001, APAC noted that while pharmacists would be more likely to deliver CMIs if they were shorter, surveys had revealed that consumers actually wanted more information, not less (APAC, 2001). APAC's view was that CMI length in terms of content was not a problem, provided users could find the information they needed; but length in terms of number of pages was a problem—and many CMIs, using the crude templates published in 1994, had multiple pages.

One-page CMI: phase 2

In early 2002, QARG, on behalf of the Commonwealth Government, engaged CRIA to return to the issue of developing a more compact format suitable for all CMI, and to test the new format for its performance with medicine users. The pharmaceutical industry agreed to undertake the testing of any new design.

CRIA began, as before, with a word count of CMI. Three companies—MSD, Astra and Novartis—provided CMI word counts for all their medicines. Below is a summary of the main statistics.

CMI word count: MSD/Astra/Novartis
mean	1900
median	1771
99th percentile	3483
95th percentile	3014
90th percentile	2690
85th percentile	2520

The greatest number of words in the one-page Sigma designs was 2,518, suggesting that the Sigma design could deal with at least 85% of all CMI provided that the general content was placed elsewhere. But the pharmaceutical industry did not want the CMIs to be edited in any way—unlike the Sigma design where the general information had been separated from the medicine-specific—as that would have required the rewriting of every CMI; so the newly-formatted compact CMIs would have to contain exactly the same number of words in the same order as the originals.

The CMI Fonts

The Poppl fonts used in the Sigma design could not be used in this second phase because neither industry, government, nor pharmacy was prepared to buy a license to use them. An examination of the public domain fonts available at that time suggested that none had the necessary legibility characteristics and range of weights needed to work at the small sizes (6 and 7 points) necessary to fit the largest number of words onto the two sides desired by the pharmacists.

In the absence of a suitable public domain font, a new font family specifically for CMI was developed, called, not surprisingly, 'CMI'. It contained CMI Regular, CMI Light, and CMI Heading. It derived from Sassoon, a font used in primary school readers (Figure 16).

Figure 16: The CMI font family

• The CMI font family was designed to have high legibility characteristics at small and medium sizes (6pt-14pt).

• The stroke width was optimised on the CMI Regular font for laser printers operating between 300 to 600 dpi, and the CMI Light for offset printing up to 2400 dpi.

• Both a regular and italic font for CMI Light and Regular were created. This meant that the italic was a true italic with its own defined character legibility, rather than a regular font which had been slanted in order to give it an italic appearance but retaining regular characters.

• The spacing between characters was optimised to enhance overall legibility.

• Each font contained a full set of alphabet characters and other characters used in CMI.

• The font family contained a number of special new characters, unique to their application:
  – special CMI dot point
  – CMI counselling point
  – character containing a new CMI Logo.
  The technical advantage of having these three special characters as part of the font was that it reduced the file size to a minimum, requiring no extra graphics to be stored.

• The font family was manufactured in PostScript and TrueType for both Mac and PC. It could be viewed on-screen, printed, or embedded in documents. There were no technical limitations on its use across any of the major platforms and applications.

(Note that copyright of the font family is with the Communication Research Institute and the Commonwealth. This ensures that the font can be used only for those purposes deemed appropriate by the copyright holders. It also ensures that if the font family needs to be modified at some future date, it can only be modified by professionally competent typographers. In practice this means that the font can be made freely available for CMI use in Australia at no cost to the end users.)

A new CMI layout

A structure of the new 2002 CMI used the advanced document-structuring features of the 1994 Writing about medicines for people. These enabled the use of the same content but with a very different appearance. Using the word-count data, CRIA developed a design which would get approximately 3000 words on the two sides of one A4 page, thus ensuring that 95% of CMI would fit. This was done to meet one of the pharmacists' main objection to the 1994 design: that it occupied too many pages.

To maximise the use of space, while maintaining the readability and usability of the document, the design was developed using professional digital design technology—Adobe InDesign software. To retain the advantages of better layouts and appearance provided by this technology, CRIA decided that the final output should be to pdf.

InDesign's sophisticated style-sheet technology enabled considerable refinement of the existing CMI styles to maintain CMI usability and ease of production. CRIA developed 21 paragraph styles and the new character style for counselling instructions. The new fonts and styles were set up so that changes could be made to them if this was warranted by test results.

Because the pharmaceutical industry was unwilling to edit CMI in any way to meet the pharmacists' requirements, as was clear from the industry's reaction to the Sigma design, the text for this new CMI could not be edited and had to be presented as a continuous stream of text. This inevitably compromised the navigational features of the design and hence its usability. To offset this, the styles were designed to ensure that related headings, instructions and explanations were kept together in the same columns. The new design also introduced some features to enable users to differentiate more easily between instructions and explanations. This was augmented by adding some text to the introductory section—What is in this leaflet—which explained how to use the structure. To make the best use of the one-page space limit, the new design used a 4-column A4 portrait format (Figure 17).

Figure 17. Click to enlarge.

In sum, the new design was a step backwards from the innovation of the untested Sigma design, but there was some hope that the new design would perform at least as well as the 1994 design and possibly better.

The full text for the new CMIs was provided to CRIA by the pharmaceutical companies AstraZeneca, Merck Sharp & Dohme, and NOVARTIS. The CMI were for Comtan®, Imdur®, Pethidine injection BP, Rezulin® tablets, and Zocor®. We did not edit the text, but we replaced the product name (repeated in the section headings) with 'it' in accordance with the Usability Guidelines and recommended some minor changes to the text.

Once the basic design was complete, CRIA tested the process and determined the time it would take to format CMI for production.

Formatting without using styles from the start

The CMI texts were supplied to CRIA in Microsoft Word files. None of the files used styles. They used typewriter conventions (double spacing at the end of sentences, two carriage returns for paragraph spacing). This meant that each document had to be manually fitted into the new layout, with each paragraph being assigned a style (Instruction style, Heading 1 style, and so on) and a number of search-and-replace routines had to be applied to eliminate unnecessary spacing and paragraph breaks.

This operation required approximately one hour of a skilled technician's time for each CMI. As well, because the process was manual, there was a high probability of formatting and literal errors at this stage.

Formatting using styles from the start

As a trial, CRIA set up a CMI in Microsoft Word with the appropriate styles and then imported it to an InDesign template. Minor changes to assigned styles were applied to ensure that the copy made best use of the column structure, along with some topping and tailing (inserting CMI logos and page directions), and then the file was exported to a pdf format. The entire operation took less than five minutes. In a production environment one would probably introduce a number of quality control checks to ensure the integrity of the process, which would add to this time. However, it was an indication of how automated the system could become.

The critical factor was the consistent allocation of styles to the primary text at the point of authoring rather than at a later stage, greatly speeding up the production process (five minutes versus one hour per CMI) and ensuring the integrity of the document throughout the production process.

Production

During the development of the new design we sent copies of the CMIs to the companies for proofing. This was done as soon as styles were assigned to each paragraph, but before the actual format of the styles had been finalised. As anticipated, most of the corrections were in the application of styles to specific paragraphs. We received some critical comments about the specific appearance of the styles; most of these comments anticipated the final layout.

The final formatting of the styles and layout were developed in Adobe InDesign 1.5.2. This software was chosen for many reasons: styles assigned in Microsoft Word are maintained; it gives the best control of text layout and styles in current high-end desktop publishing software; it provides excellent control and compatibility with all major importing and exporting formats; documents can be exported as tagged text and in a variety of other formats such as rtf, xml, and html; and in particular, it provides excellent control over pdf production.

The final designs were exported to pdf files. The file sizes ranged from approximately 150K for Pethidine to 250K for Zocor.

CMI for testing

Figures 18 to 21 show two of the reformatted CMI that were used in the testing.

Figures 18 to 21. Click to enlarge.

On 15 March 2002, under the auspices of EDWG, pharmaceutical company representatives tested the newly formatted CMI for Comtan (Novartis), Imdur (AstraZeneca), Pethidine (AstraZeneca), Rezulin (Pfizer) and Zocor (Merck Sharp & Dohme). Testing was conducted at AstraZeneca.

The CMIs were tested for two purposes:

• to find out if the revised format improved the CMI's performance in providing medicine information, compared to the performance of the then current multiple-page 3-column format: participants were asked questions about using the medicines to see how easily they could find and use the information on the revised CMI
• to see if users found the new format acceptable: participants were invited to comment on its good and bad aspects.

Test participants

Participants comprised 13 volunteers from the Diabetes Association of NSW, Ryde branch, ranging in age from 58 to 82.

Method

The method followed the directions for diagnostic testing of CMI in Writing about medicines for people (Sless & Shrensky 2006)

Participants were given the CMI to read at leisure. When they felt ready, they were asked a series of 10 or 11 questions and then asked for general comments on the good and bad points of the CMI. At the end of the session they were shown the same CMI in the current (1994 design) 3-column format and asked to comment on their preference, if any.

The results were compared with the results of testing the CMI in its 3-column format, using the identical questionnaires administered to the 49 test participants in 1994. See the individual test results in Appendix 1.

Summary of testing results

Overall the new format CMI tested almost as well as the current format, although there were indications that it could be further improved: participants tended not to read to the bottom of columns, so were unable to find information in this position; and they were also less likely to use subheadings to find information, owing to the subheadings' lack of prominence and/or differentiation from other text in the new format.

Although all participants could read the new font, many of them commented that it was very or too small and would be hard for 'older' people to read.

When asked to briefly compare the new format CMI with the current CMI, the majority of participants preferred the current CMI, mainly because of the larger font size, and they also favourably commented upon the formatting, spacing, and headings. Altogether, the current CMI was 'easier' and 'more comfortable' to read.

Bill Willcox, CHF representative on EDWG, who attended the testing, reported that recurring themes discussed by the participants waiting for interviews were:

• surprise that they did not know of the existence of CMI (as members of a consumer group active in the consumer movement); and
• observations that they were generally treated paternalistically by their doctors and pharmacists on the basis of "don't worry - I know what's best for you".

Conclusions of the testing

Overall, the new format CMI tested well against the current format. But participants expressed disquiet. The majority preferred the older format CMI mainly because of its larger font size; although they could read the new font in the revised version, many commented that the small size would make it hard for older people to read. Many also commented that the spacing and headings in older CMIs were easier and more comfortable to read. In the revised format, by contrast, subheadings were not as clearly differentiated from the rest of the text, so participants often failed to use them to locate information owing to their lack of prominence. Participants also tended not to read to the bottom of columns and so were sometimes unable to locate information in this position.

One-page CMI: the final phase

After considering the results of testing, EDWG recommended that subheadings should be made more prominent to help users locate information and that there should more differentiation in font format, following which it would be appropriate to retest the new CMI layout. Such refinements were well within the range of what was possible to adjust after a first round of testing. Indeed, with a new font family and a new layout these were exactly the kinds of faults to be expected.

But it was clear that both consumer groups and industry were reluctant to embrace a design that consumers compared unfavourably with the existing design and where further work was needed.

The Pharmacy Guild made little comment on the test results. Their principal concern remained with the number of pages and with a new concern over the file size. As noted above, the file size of the final PDF outputs would be between 150k and 250k. This was about 10 times the current file size. It was pointed out that this might require pharmacists to install larger hard drives in their computers. Realising this might involve a cost, the Guild objected, and without their support no further work was done on the project.

The concept of counselling points was also abandoned because there was no agreement on which authority, and on what basis, a decision could be made on the positioning of the points.

So in the end, technical and organisational issues remained unresolved, and the project to develop a new CMI format was shelved.

Conclusion

It has been known for some time that many problems impede the automatic provision of CMI: for example, medicine users often being unaware that CMI is available, and healthcare professionals lacking awareness and training. One important barrier is that community pharmacists resist automatic provision of CMIs; even though pharmacists have, since 2000, been provided with readiness funds for printing and financial incentives for counselling CMIs, they still claim that CMIs are too long (that is, more than one double-sided A4 sheet) to hand out routinely.

To address the problem of length, CRIA designed new formats to enable the majority of CMIs to fit onto one double-sided sheet. The first design, for Sigma, put preprinted general information one side of a sheet, and medicine-specific information on the second side. The second design, for QARG and the Commonwealth Government, was unable to use the previous strategy as it would have meant that all existing CMIs would have to be rewritten, so a completely new format was devised. Test results demonstrated that clarity was sacrificed—a large amount of information into a smaller space doesn't go easily without some loss. Refinements would have meant slightly increasing the length of CMI so that only 90% would fit onto two pages, and some increase in file sizes, a result unacceptable to pharmacists. The potential, though modest, added cost of larger hard drives was used by the pharmacists to veto any further development.

Thus the crude CMI format first designed in 1994 is still with us.

Despite these setbacks, the development of CMI in Australia and the usability of CMI for consumers is a high point in this field internationally. Though CMI seriously lag behind best practice in information design, they do provide usable information for consumers. The main obstacle to improved CMI design and distribution lies with the outmoded technology used in pharmacies at the point of sale and the reluctance of pharmacists to give CMI to consumers.

At the time of completing this paper there was no indication that pharmacists had changed their position. Effectively they block the road to improved CMI. The conclusion must be that it is time to rethink the mode of distribution and delivery of CMI in Australia. As Figures 5 and 6 show, it is possible outside the constraints of pharmacy technology to develop CMI to a high standard. In the public interest, it is important that CMI be brought up to the best practice standard that apply in other areas of document design.

With sophisticated information design practice, the methods and skills applied to designing hard copy CMI can be applied to internet delivery. Indeed, with existing style functions and xml or similar technology, it would be possible to enable consumers to tailor CMI documents for their own needs, enabling the CMI to meet far more of criteria of good document design. Moreover, this could be achieved without industry having to change the existing content, thus eluding the recalcitrant pharmacists and their outmoded technology.

A reasonable objective for future CMI

bin-avoidance criteria	Pack insert and internet cmi	why?
1. credible	yes	manufacturer's brand in pack insert, and authoritative source on internet
2. respectful	yes	greater flexibility in presentation
3. attractive	yes	greater flexibility in presentation
4. physically appropriate	partially	restricted by packaging limitations and internet access technology
5. socially appropriate	partially	some control with print fully customisable on the internet by consumers to meet their specific needs
criteria for reading and using
6. usable	yes	tested
7. efficient	yes	greater flexibility in presentation
8. productive	possibly	subject to monitoring in use

On this basis, we believe that it is time to abandon any further development of CMI for distribution by pharmacies, return to pack inserts, and where appropriate, embrace the commonly available technology of the internet.

References and further reading

Baume Peter 1991
A question of balance: report on the future of drug evaluation in Australia. Canberra: Australian Government Publishing Service, 1991.

Graham Janne D 2008
What to do about CMI. Health Voices Forum (newsletter of the Consumer Health Forum of Australia

PSA 2007
Significant changes to CMI guidelines
Media release, 21/2/2007 [retrieved February 2009]

QUM 2003
Measurement of the quality use of medicines component of Australia's National Medicines Policy: Second report of national indicators. Department of Health and Ageing.

Shrensky R & Sless D 2008
Choosing the right method for testing.

Sless D & Wiseman R 1994
Writing about medicines for people: Usability guidelines and glossary for Consumer Product Information.
Canberra: Department of Health and Family Services.

Sless D & Wiseman R 1997
Writing about medicines for people: Usability guidelines for Consumer Medicine Information 2nd edition.
Canberra: Department of Health and Family Services.

Penman R Sless D & Wiseman R 1996
Best practice in accessible documents in the private sector.
in Putting it plainly: Current developments and needs in plain english and accessible reading materials.
Canberra: Australian Language and Literacy Council

Sless D 2004
Designing public documents
Information Design Journal + Document Design Journal 12 (1) 24-35

PHARM 1999
Using Consumer Medicine Information (CMI)
A guide for consumers and health professionals retrieved 31 May 2009

Appendix: Testing the 4-column CMI

These reports were provided by Heather Friedel, Jenny Chu Marlene Arens, Susan Parker and Sylvia Roins. They are presented here, lightly edited to achieve some consistency between them. Otherwise they are the reports that were presented to APAC as part of this project.

Comtan CMI

Results

Key: Y = found and understood; X = not found; +/- not understood

Discussion

The results were compared with results of testing the 3-column Comtan CMI using the identical questionnaire administered to 10 test participants (aged 22–76 years) at Sydney Adventist Hospital on 21-22 September 1999.

The new CMI format tested as well as the 3-column format in current use. There were no complaints about the 4-page layout or the length of the CMI. All participants liked the major headings being in white on a dark background and made use of these headings in locating information.

There were some problems with locating information, which seemed to happen most often with text located at the bottom of a page; the poor response to question 2 was due more to the fact that the heading Taking Other Medicines was at the bottom of column 2, and the instruction on taking an MAOI with Comtan was also at the bottom of column 2 with the explanation at the top of column 3.

With the current 3-column format, care is taken not to have important headings at the bottom of a page and not to split instructions and explanations. With the new format, the need to limit space meant that this convention was sometimes disregarded.

Consumer Preferences

When shown the current 3-column CMI at the end of the interview, 4 of 6 participants preferred it to the new format and the other 2 had no preference. Although all 6 could read the new version and find the information in it, all 6 participants commented that the font was too small and that the 3-column version was much 'easier' and 'more comfortable' to read. They felt that instructions and explanations in the 3-column format were much more clearly differentiated using bold and plain text and that they should be more clearly differentiated in the new format through the use of differing font size or type. They felt that some of the subheadings could also be larger or better defined. Participants also wanted to know why there were markers against some of the text. When told that these were counseling points' for the pharmacist to use, participants suggested that the icon and a statement explaining their purpose should be part of the general explanation at the beginning of the CMI.

Conclusion

The performance of the new format CMI is acceptable. However, consideration should be given to (1) more clearly differentiating between instructions and explanations, (2) allowing for movement of text so that subheadings do not appear at the bottom of a page and instructions and explanations are not separated, and (3) adding the counseling point icon and description to the general instructions on how to read the CMI.

Imdur CMI

The Imdur CMI fits on 3 sides in the current format, and on 2 sides in the proposed format.

Results

test question	PC 59 yo M	AR 74 yo M	BM 66 yo M	DJ 68 yo M	DG 79 yo M
	Y Very easily	Y Very easily	Y Very easily	Y Very easily	Y Very easily
2. Usual dose	Y Very easily	Y Very easily	Y Very easily	Y Easily	Y Easily
3. Two a day	Y Very easily	Y Very easily	Y Very easily	Y Easily	Y Very easily
4. Miss a dose	Y Very easily	Y Very easily	Y Very easily	Y Very easily	Y Easily
5. What if not taken regularly	Y Some difficulty	Y Some difficulty	X Not found	Y Some difficulty	Y Some difficulty
6. Drink alcohol	Y Easily	Y Easily	Y Very easily	Y Easily	Y Easily
7. Most common side effect	Y Some difficulty	Y Easily	Y Very easily	Y Very easily	Y Very easily
8. What if headache	Y Very easily	Y Very easily	Y Very easily	Y Very easily	Y Very easily
9. Overdose	Y Easily	Y Easily	Y Very easily	Y Very easily	Y Very easily
10. When stop taking	Y Very easily	Y Some difficulty	Y Very easily	Y Some difficulty	Y Some difficulty
11. Storage	Y Some difficulty	Y Easily	Y Very easily	Y Easily	Y Very easily

Key: Y = found and understood; X = not found; +/- not understood

Discussion

The results were compared with results of testing the Imdur CMI in 3-column format using the identical questionnaire administered to 9 test participants (aged > 55 years) of the Greenacre branch of the NSW Pensioners and Superannuands Association on October 14, 1994. The results of testing with the proposed format CMI were similar to the results for the 3-column CMI tested in 1994.

Subjects had difficulty finding the answers to two questions, in particular:

• Question 5, which was the only answer presented in the smaller type size used for additional information,
• Question 10, which was at the very top of the right hand column of the CMI.

Despite scanning the page, whether using headings or looking up and down columns, most participants skipped over these two pieces of information more than once. In the case of question 5, participants commented that this was because of the small type size. The reason for the difficulty in locating the answer to question 10 is less obvious; it may have been because of its slightly unusual position or its separation from the heading Things You Must Not Do.

Three of the participants commented that the type used for instructions did not stand out sufficiently from the additional information.

Most participants used the white on black headings to navigate the document, and three participants commented that these headings were helpful and that subheadings should also be more prominent.

When answering question 9, about what to do in case of an overdose, most participants commented that either the contact details for the Poisons Information Centre or the Overdose heading should be given more prominence.

Some participants commented that the second page of the CMI was wasted because it contained only one column of information, and suggested that the type size be increased to make use of the empty space and make the document easier to read.

One participant commented on the markers for counseling points, and suggested that their meaning be explained at the beginning of the CMI.

All participants could carry out the instructions.

Preferences

Four of the five participants preferred the current 3-column format, primarily because of its larger font size. One participant mentioned that he preferred the style of list bullet used on the current CMI to the arrows on the proposed format. Subjects also commented that the bold instructions on the current CMI were easier to distinguish from the additional information.

The participant who preferred the new format said that the length of the current CMI—3 pages compared to less than 2—was a bit daunting.

Good and bad points

Subjects' additional comments about good and bad points of the CMI are given below.

PC, 59

“Very tiny writing”. “Nothing is easy to find”. “My ninety-year-old mother couldn't?t read it”. Column layout good because easier to scan, but there needs to be more white space between the columns to make it easier to read. Standardised icons (eg for use in pregnancy, use by children, with alcohol, driving, etc.) would be useful to point readers towards the relevant sections. Highlighting of instructions could be made more prominent. Poisons Information Centre details should be a standard footer on all CMI so that it is easy to find in a hurry. English is good, simple and straightforward. Only italic instructions should be included in CMI, and additional information should be removed to keep it short.

Comparison of current to new: Preferred current format CMI because type bigger and more white space between columns, but would prefer white on black headings.

AR, 74

Two different sizes of fonts aren?t sufficiently different and need to be distinguished more. What is magnifying glass symbol? Needs to be explained. Every “not” should be twice the size and in capitals. Some other headings (eg Overdose) should be in white on black format. Overall, type is too small, but once you?ve read it, it seems OK. White on black headings stand out well.

Comparison of current to new:Current CMI format easier to read because bigger type, and the difference between the two font types is more obvious. Prefer it because it is more readable. Bullets are good for lists (compared to arrows on new format, which ”don?t mean anything“.)

BM, 66

Nothing particularly difficult about layout. ”Well set out“. Should make print bigger to fill up two sheets instead of having wasted space on second page. Can read clearly. Overdose heading needs to be more prominent because this is important information. Should add web site address to leaflet or box, because participants always look up medications on the internet.

Comparison of current to new: Larger print is easier to read and headings are clear. ”A compromise of print size between the two would be better and should print on both sides of the page.“

DJ, 68

”Awful lot of information“. ”Easy to skip over“. Overdose heading and Poisons Information Centre details should stand out. Difficult to find information in a hurry.

Comparison of current to new: Prefers current format CMI because type bigger and ”less likely to skip over, more likely to read and absorb“, but Poisons Information Centre details still do not stand out enough. ”Length doesn't matter if you are getting the message across“. Proposed format CMI looks like more information and that the information is less important.

DG, 79

Instructions need to be larger, more prominent than additional information. Headings stand out well and cover topics well. Type is a bit small but larger type will make CMI longer. Could type size be increased to fill two pages rather than having wasted space?

Comparison of current to new: 3 pages are ”a bit daunting“. Emphasis in bold is easier to pick out than in the new format. Pretty good, but maybe should use up spacing at ends of columns more and make type slightly smaller to fit onto 2 pages.

Conclusion

The proposed CMI format performed as well as the current 3-column format, although most participants preferred the current format. Shortcomings of the new format were the small type size (especially for additional information, which some participants tended to skip over), and the lack of distinction between instructions and additional information. Most participants preferred the black on white heading style. participants' concerns about quickly locating important information in the CMI were reflected in their suggestions for more prominent subheadings, particularly for the overdose section.

NB The pdf file for Imdur did not print correctly on AstraZeneca printers.

Pethidine CMI

Results

key: Y = found and understood, X = not found, +/- not understood

discussion

The results were compared with results of testing the Pethidine CMI in 3-column format using the identical questionnaire administered to 13 test participants (aged 49 to 89 years) at St Vincents Hospital and Concord Hospital over a period from December 1994 to January 1996.

Compared with the previous testing of the Pethidine CMI in the 3-column format, the results were slightly worse in terms of ease of finding and understanding the information. Three of the four test participants had difficulty locating information at the bottom of the page, seeming to stop reading about an inch from the bottom of the text. This is reflected in the poorer test results, since the answers to several of the poorly answered questions were at the bottom of a column.

Everyone could read the new format; but only one participant (CC, aged 59) preferred it to the current 3-column version, because it would use less paper and may be more likely to be handed out.

No participant commented on the symbol for counseling points.

good and bad points

The individual comments about the good and bad points of the leaflet are listed below.

JR, 72

Good, only it's too small – an older person couldn't read it. The sub-headings could be bigger. It's informative, has everything you want to know.

Current format:

That's great – anyone could read it. If you can't see it, it's not much use. Very easy.

CC, 59

Should use the same word – either use or take (explained that “use” was for medicines someone else usually administered), agreed OK. Very helpful, would use the CMI not necessarily just for side effects. Clearly set out, wouldn't miss anything, quite well done. 3 weeks ago would have found it difficult to read (vision has improved since diagnosis and treatment of diabetes).

Current format:

Uses more paper. Easier to read for elderly people.

VM, 69

Good, this information needs to be in writing. Need to sit down and read it. Suggested a different order of information, but agreed it was OK when the rationale was explained. Needs more concentration and plenty of time to read it.

Current format:

Likes typing better – bigger. Much easier for the elderly. Pharmacists more likely to print the new version.

SH, 82

Expiry date a good idea – patients need to know. Excellent.

Summary

There was universal approval for the information contained in the CMI. The 4 test participants (all of whom wore glasses to read) could read the document, but there was a preference for a larger font size.

NB: The pdf file for Pethidine did not print correctly on AstraZeneca printers.

Rezulin CMI

Results

(Y= correct answer given)

Discussion

All thought the instructions were OK. Subjects showed varied ability in locating information. 4 out of 5 participants expressed concern about small print size, and 1 participant was concerned about making sub-headings and important words stand out. There were also comments about expression and choice of words. All participants preferred current format for clarity.

Good and bad points

• Good headings, well prepared, all information is there (CEB).
• Black print on white paper (ie no colour) is the best according to the Royal Blind Society so this is good as it is easy for people to read. Black print on yellow paper is the second best (LL).
• Well set out, headings are good (JM).
• Clear, white on black headings are good (DL).
• Concise – 2 pages, and headings are clear and understandable (CC).
• Very small print (CEB).
• Print is too small. The Royal Blind Society recommends nothing below 12 point for the general public (LL).
• Need heavier type (bold) for sub-headings (LL).
• Need important words, eg pregnancy, in bold print or something to stand out (LL).
• Small print (JM).
• Prefers larger font but personally not that concerned with font size. Others however may require glasses or magnifying glasses for small print (DL).
• Small print – 3 weeks ago I would not have been able to read this as I was diagnosed with diabetes (CC).

other comments

• Larger print size is better (CEB).
• ”Product Description“ should be after ”What is it used for“ so people know what they are taking (LL)
• Use of the word ”products“ rather than ”medicines“ in section ”any other medicines – including any which you have bought from a pharmacy, supermarket or health food store“ as people do not consider vitamins or natural products as medicines (LL).
• ”Take to pharmacy for disposal“ rather than ”ask your pharmacist what to do with any tablets left over“ – ie it gives an instruction of what to do (LL).
• Use of ”Take one tablet a day“ rather than ”Take one tablet once a day“ as the word ”once“ means eleven in Spanish and there is a risk of overdose (LL).
• Should clearly state "with glass of water" under How To Take It, as it states ”Take Rezulin tablets once a day with food“ which implies with food rather than water (DL).

Preferences

All preferred the current format because of larger print and headings. The main points are more obvious in the current format (LL).

Zocor CMI

Results

* PhD in Microbiology & Chemical Engineering

(Y= correct answer given)

Discussion

4 out of 5 participants were not able to locate information about addiction probably because of its location. There was also some confusion about ”Other uses“ and ”Side effects“. All participant found that the print size in revised format was too small or that they could read the 3-column CMI better.

Good and Bad points

• English is good and simple (PETER).
• Good coverage of all aspects (PETER).
• Liked white heading on black (PETER).
• Format for side effects good, although could make them more noticeable (JR).
• Good document (CB).
• Writing is alright (JR).
• Print size is all right (VM).
• Writing is very small - need to present it in a way so eye comes quickly to it (PETER).

• More bolding & highlighted required (PETER).
• Writing a bit small for older people, it could be bigger (like the font size of the section headings) (JR).
• Some headings don't stand out (LORRAINE).
• Headings under sections could be larger/bolder (LORRAINE).
• Terrible print size – according to the Royal Blind society, font size should not be < 12 points (LORRAINE).
• Print size is too small (CB).

Comments and Suggestions

• Headings of different CMI should be standardized (PETER).
• Suggest use of icons/pictures that are uniform throughout the industry (PETER).
• Did not know what the magnifying glass was, although it did bring his attention to the text next to it (PETER).
• Side effects listed as bullet points is good, makes them easy to find (LORRAINE).
• In ”Taking other medicines“ section, recommend adding vitamins and herbal products as things to tell your Dr about as people often don”t think of these as medicines (LORRAINE).

pereference

• Prefers 3-column CMI as bigger print, not crowded, and can read column easily (more white space), and page numbers not a problem (PETER).
• 3-column CMI is better – I would go straight to reading the bolded text as the heavy print is obviously important. Also prefer the section headings in the 3-column CMI (LORRAINE).
• Prefer 3-column CMI (CB).
• Prefers 3-column CMI 100% as its very easy to read. Small print may turn people off reading it (JR).
• Can read 3-column CMI better (VM).

Observations during testing

• The participants did not seem to notice/use the headings under the sections.
• The participants did not seem to notice any difference between the italicised font (ie. the instructions, which are normally bolded in the 3-column CMI) and the normal font.

Conclusions

Everyone commented that the print size of the new 1-page CMI for ZOCOR was too small, and preferred the current 3-column format over the new 1-page CMI.