<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-0905">
  <Title>Tailored Patient Information: Some Issues and Questions</Title>
  <Section position="7" start_page="31" end_page="31" type="concl">
    <SectionTitle>
4 Conclusions
</SectionTitle>
    <Paragraph position="0"> TPI systems are likely to be of increasing interest to health care providers. They clearly work to some degree, and they should become more effective as they start using more advanced technology, such as NLG. However, it is not sufficient for a TPI system to be clinically effective in order to be fieldable; it also needs to be cost-effective, acceptable to individual users (patients as well as medical practitioners), have low data-entry costs, and incorporate a satisfactory solution to the mistakes issue. This is a daunting set of requirements, and may explain why although many TPI systems have been developed in the lab, few have been fielded.</Paragraph>
    <Paragraph position="1"> We hope that a better understanding of these issues will help TPI developers (including ourselves) produce systems that are more likely to be deployed and used in the real world. We believe that TPI technology has the potential to make a real impact on health, especially given the increasing importance of life-style and compliance issues; good health is mostly a function of actions and decisions taken by patients, not by health-care professionals.</Paragraph>
    <Paragraph position="2"> But this potential will only be realised if we can build systems that are not only technologically ingenious and clinically effective, but also are easy to deploy and use.</Paragraph>
    <Paragraph position="3"> We would like to conclude by saying that we believe that these fielding problems will decrease in the future. In particular, cost-effectiveness should increase as technology improves; acceptance among health-care professionals should become easier as more such people become computer literate and friendly; data-entry should become less of a problem as PRS systems become richer and more standardised; and people may become more tolerant of computer mistakes if they adopt the &amp;quot;make mistakes less often than a doctor&amp;quot; criteria. So, in ten years time it should be much easier to deploy a TPI system; all the more reason for researchers to work today on developing appropriate technology and identifying good applications.</Paragraph>
  </Section>
class="xml-element"></Paper>