Research Paper n
Assisting Consumer Health Information Retrieval with
QING T. ZENG, PHD, JONATHAN CROWELL, MS, ROBERT M. PLOVNICK, MD, EUNJUNG KIM, MS,
LONG NGO, PHD, EMILY DIBBLE, PHD
A b s t r a c t
millions of people, many of whom have problems forming effective queries. We have developed and evaluated a tool to
assist people in health-related query formation.
Objective: Health information retrieval (HIR) on the Internet has become an important practice for
Design: We developed the Health Information Query Assistant (HIQuA) system. The system suggests alternative/
additional query terms related to the user’s initial query that can be used as building blocks to construct a better, more
specific query. The recommended terms are selected according to their semantic distance from the original query, which
is calculated on the basis of concept co-occurrences in medical literature and log data as well as semantic relations in
Measurements: An evaluation of the HIQuA system was conducted and a total of 213 subjects participated in the
study. The subjects were randomized into 2 groups. One group was given query recommendations and the other was
not. Each subject performed HIR for both a predefined and a self-defined task.
Results: The study showed that providing HIQuA recommendations resulted in statistically significantly higher rates
of successful queries (odds ratio 5 1.66, 95% confidence interval 5 1.16–2.38), although no statistically significant
impact on user satisfaction or the users’ ability to accomplish the predefined retrieval task was found.
Conclusion: Providing semantic-distance-based query recommendations can help consumers with query formation
j J Am Med Inform Assoc. 2006;13:80–90. DOI 10.1197/jamia.M1820.
Health information retrieval (HIR) on the Internet has become
a common and important practice for millions of people.1
Health consumers of varying backgrounds perform HIR for
themselves as well as for friends and family, and to merely
satisfy their own curiosity, as well as to make medical deci-
sions. Because of the vast amount of information available
and the adhoc natureofinformation gathering by consumers,
HIR is not always efficient and effective.
Query formation is a major aspect of consumer HIR that is in
need of improvement. One observation study has shown that
consumers’ HIR queries tend to be too short and general.2
Although current search engines are fairly good at retrieving
appropriate information, they still depend on the queries to
set the correct retrieval goal. If queries do not reflect users’
specific information needs, they will lead to results that do
not address those information needs. For instance, we once
interviewed a user who wanted to know ‘‘Are there natural
substitutes for the hormone replacement therapy Prempro?’’
One of the queries this person typed in was ‘‘natural hrt.’’ It
was not surprising that the query failed to yield the correct
Internet queries tend to be short regardless of the search do-
main: users do not type more than 2 or 3 query terms on av-
erage.3,4When searching for health information though,
many consumers’ limited knowledge of medical vocabulary
contributes to the construction of simplistic queries. For in-
stance, when a consumer we interviewed could not remem-
ber the exact name of a drug, he had to use the more
general query term ‘‘antidepressant.’’
To help consumers better articulate their health information
needs, we have developed and evaluated a novel system,
the Health Information Query Assistant (HIQuA), to recom-
mend alternative/additional query terms. The recommended
terms are deemed to be closely related to the initial query and
can be used as building blocks to construct more accurate and
specific queries. By relying on user recognition instead of re-
call, our tool attempts to make query formation easier.
Background and Significance
Consumer Health Queries
We have interviewed consumers and analyzed log data of
health-related consumer queries in some of our previous
work. Three findings from our previous studies are: (1)
Affiliations of the authors: Decision Systems Group, Brigham and
Women’s Hospital, Harvard Medical School, Boston, MA (QTZ, JC,
RMP, EK); the Department of Biostatistics, Harvard School of Public
Health, Boston, MA (LN); Bunker Hill Community College, Boston,
Supported by National Institutes of Health grant R01 LM07222.
The authors thank Lowell Ling and Chantal Friedman for assisting
with the evaluation study.
Correspondence and reprints: Qing T. Zeng, PhD, Department of Ra-
diology, Decision Systems Group, Thorn 309, Brigham and Women’s
Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115;
Received for review: 03/04/05; accepted for publication: 08/27/05.
ZENG ET AL., HIR Query Recommendations
consumer queries are short (usually no more than 1 to 2
words on average),5(2) most terms in consumer queries can
be mapped to concepts in medical vocabularies,5and (3) the
terms and concepts consumers use often do not accurately re-
flect their information needs and do not form effective
queries.2,6The problem of overly general queries and ineffec-
tive search strategies in consumer HIR has also been reported
by Eysenbach et al.7
In HIR, many queries are not only short, but also not specific
enough to describe the information needs. One reason is that
consumers, unlike clinicians or research scientists, have lim-
ited knowledge of medicine. As a result, they requiremore as-
sistance in query construction.
The work by Fredin et al. also sheds some light on this issue.8
They suggest that Internet information retrieval is an iterative
process, during which the information retrieval goals are con-
stantly refined and revised. Consequently, queries need to be
refined and revised. Our system, if successful, can make the
process of refinement and revision more convenient for users.
Researchers have developed many techniques to improve in-
sion, i.e., adding additional terms to the original query.9
documents themselves. A thesaurus may offer synonyms,
antonyms,descendents, orother related terms.Retrieval feed-
back methods analyze the ‘‘best’’ returned documents, as
determined by the user or by some ranking algorithm. Co-
occurrence data of the query and other terms in certain data
sets, for instance, log data that records the search behavior of
previous users, have also become sources for expansion
to the original query. In interactive systems, related terms are
terms to expand or replace the original queries.
In recent literature, variations of the basic query expansion
techniques have been reported. Some techniques combine dif-
ferent expansion methods, for example, combining retrieval
feedback with co-occurrence information13or combining sev-
eral thesauri.14Some have explored the fuzzy nature of relat-
edness between terms or concepts.15–17The fact that Web
users are often not good at constructing queries has led to
more studies on interactive methods, while the availability
of large query logs from Web sites has provided a rich source
for mining term and concept relations.18–21
In biomedical informatics, there have been a number of appli-
cations that have used query expansion techniques for search-
ing literature.22–27The sources of expansion terms have been
medical vocabularies, retrieval feedback, and co-occurrence
data. A set of methods has also been developed to transform
natural language questions or queries into computer-friendly
representations such as Boolean expressions or conceptual
graphs.28–31(Similar studies have been carried out on search-
ing patient medical records,32,33which we will not elaborate
For retrieving consumer health content, previous studies
have explored query expansion, reformulation, and sugges-
tion. For example, the study of Gobel et al.34added broader
and narrower concepts automatically to user queries accord-
ing to entries in the MeSH Thesaurus. McCray and
colleagues’ study35utilized a variety of strategies such as syn-
onym expansion, spelling correction, and suggesting more
general queries when no results are found, among others.
Finally, we have conducted experiments,36as have Patrick
et al.,37that examined the impact of reformulating consumer
queries with professional synonyms. Of all the studies men-
tioned, however, none explored concept relations beyond
synonymy or hierarchy.
When dealing with consumer HIR, the main query expansion
approaches have pros and cons. Automated expansion that is
based on a thesaurus or on co-occurrence data does not put
any extra burden on the user; however, it can end up being
even less effective than the original query if the original query
does not represent the user’s search goal well. Retrieval feed-
back methods suffer from the same problem even though
they may rely on some user participation before automati-
cally generating the new query. The strength of the retrieval
feedback method is its ability to learn from examples.
Query suggestion methods, on the other hand, require greater
user participation, which can be viewed as extra work for
users, but thebenefit of these methods is that evenif theinitial
query is poorly constructed, the user is empowered to articu-
late his or her needs and refine his or her queries. This article
describes a query suggestion method.
For identifying related terms to suggest to users, we consid-
ered several sources that have been exploited by previous
1. Usage patterns of consumers: Forming recommendations
from consumers’ usage patterns has the advantage of
reflecting the semantic distance among concepts in the
consumers’ mental models. The downside is that it also
relies on the extent of the consumers’ knowledge and their
recall abilities, which could be quite limited.
2. Controlled medical vocabularies: In medicine there is a
great wealth of available semantic knowledge embedded
in controlled vocabularies, so making thesaurus-based
suggestions is feasible and common. The disadvantage of
relying on a thesaurus is that it may sometimes lead to rec-
ommendations that are unrecognizable to a consumer.
and hierarchical relations. For instance, ‘‘pneumonia’’ is an
‘‘infectious disease’’ and is a ‘‘lung disorder.’’ These are im-
tions (e.g., the relations between medications and diseases)
are not extensively included in medical vocabularies.
3. Concept co-occurrence in medical literature: This provides
another source to estimate the semantic relatedness of con-
cepts. Medical literature reflects up-to-date knowledge in
the health domain. Past research has shown that a high fre-
quency of concepts co-occurring in literature is a decent
indicator of a close semantic distance between them.38,39
Generally speaking, its coverage of relations is more com-
prehensive than manually constructed vocabularies, but
To provide HIR users with recommendations that reflect their
mental models while avoiding being limited by users’ recall
abilities, we decided to combine these 3 sources. As some
other research has done, our method treats semantic distance
between concepts as a fuzzy concept.15–17Our method for
estimating semantic distance and combining sources was
Journal of the American Medical Informatics AssociationVolume 13Number 1 Jan / Feb 2006
designed specifically for the consumer HIR context and con-
sequently differs from other published methods.
The main function of the system is to identify medical con-
cepts that are semantically related to a user’s initial query
and recommend them to the user. The semantic distance
among concepts is calculated based on their co-occurrence
frequency in query log data and in medical literature, and
on known semantic relationships in the medical domain.
Topic-specific modifiers are also recommended for concepts
of several common semantic types. In addition, the system
continuously learns from user selections in order to improve
future performance. Figure 1 shows the overall design of the
Distance-Based Query Recommendations
To provide a query recommendation, the system first maps
the query into 1 or more concepts and then identifies concepts
that are related to those concepts.
Mapping to Concepts
In HIQuA, semantic relations and semantic distances exist
among concepts, not character strings. Query strings are
thus first mapped to concepts, which are defined by the
Unified Medical Language System (UMLS).40Each initial
query may be mapped to 1 or more concepts.
If the entire query string cannot be mapped to one UMLS con-
cept, HIQuA attempts to find concepts with names that
match the longest possible substrings of submitted search
terms. The search string ‘‘thrombosis attack coronary,’’ for in-
stance, will return two concepts named ‘‘Thrombosis,’’ and
‘‘Heart Attack.’’ (‘‘Heart Attack’’ is the preferred name in
the UMLS for the concept to which the string ‘‘attack coro-
On the other hand, a single string may sometimes not only
match a concept without being broken into substrings, but
can even match more than one concept. The word ‘‘cancer,’’
for instance, maps to two UMLS concepts: ‘‘Malignant
Neoplasms’’ and ‘‘Cancer Genus.’’ (In biological taxonomy,
‘‘cancer’’ is a genus of rock crabs.) Of these two, ‘‘Malignant
Neoplasms’’ is clearly the more appropriate concept to match
to in our application. Because queries are short and provide
little context for disambiguation, we are only able to disam-
biguate between concepts based on the following factors:
(1) whether the matched term is considered a suppressible
name for the concept by the UMLS (according to the
UMLA FAQ, certain names are ‘‘suppressible’’ if they have
‘‘invalid face meanings or are otherwise problematic’’
the editing distance (i.e., the number of editing operations—
deletions, insertions, and substitutions—necessary to make
two strings identical) between the term and the preferred
name of the concept (the shorter the better); (3) the number
of vocabulary sources containing the concept (suggesting a
common rather than a rare semantic); and (4) whether the
term is marked in UMLS with ‘‘,1.’’ (indicating that it is
the primary meaning of a term). The limitations of our
mapping technique are discussed in the Limitations section.
Identifying Related Concepts
The recommended concepts should be related to the initial
query concept(s); in other words, they should have a short se-
mantic distance from the initial concept(s). For estimating
semantic distance we used three sources: (1) the semantic re-
lations of concepts in medical vocabularies, (2) co-occurrence
of concepts in consumer HIR sessions, and (3) co-occurrence
of concepts in medical literature.
Medical vocabularies are a reliable source of known semantic
relations between concepts because they have been con-
structed and reviewed by domain experts. We used the
UMLS Metathesaurus relationship (MRREL) table as our
medical vocabulary source.
To complement the medical vocabularies, we used co-occur-
rence data of concepts in medical literature. The UMLS
Metathesaurus co-occurrence (MRCOC) table was used as
our literature co-occurrence source.
The third source of semantic distance is the co-occurrence of
concepts in consumer HIR sessions. The underlying relations
among these co-occurring concepts could be co-occurring
symptoms of a disease, a symptom and its location, a medica-
tion and a disease for which it has been prescribed, or some-
what obscure connections, such as the relationship between
‘‘diet’’ and ‘‘food allergies.’’ Of course concept co-occurrence
in search sessions could be incidental, but a high frequency of
co-occurrence is unlikely to be due to chance. The query log of
a consumer health information Web site, MedlinePlus,41was
obtained from the National Library of Medicine and used as
the co-occurrence source. This log contained 12 million
queries that we split into sessions based on the IP address
and time of the queries: queries from the same IP address
within 5 minutes of each other were considered to be in the
same session. Co-occurrence was calculated based on con-
cepts that appeared together in the same sessions.
For each concept mapped to the initial query, HIQuA extracts
related concepts from the three sources described above. The
first source, the MRREL table, lists semantic relationships
between concept, such as ‘‘parent,’’ ‘‘child,’’ ‘‘synonymous,’’
and ‘‘similar to,’’ among others. The second source, the
F i g u r e 1 .
alternative/additional query terms related to the user’s initial
query, which can be used as building blocks to construct a
Overall design of HIQuA. The system suggests
ZENG ET AL., HIR Query Recommendations
MRCOC table, lists pairs of concepts that have appeared
together in medical literature, along with the frequency of
their co-occurrence. The final source, the query log table
(QRYLOG), lists pairs of concepts that users have often per-
formed searches on in conjunction with each other. In addi-
tion, this table is continuously updated by HIQuA as it
continues to gain information about users’ search habits.
We set no limit on the number of related concepts. We did,
however, establish the following exclusion criteria in order
to eliminate inscrutable relations and accidental co-occur-
rences (see Table 1).
Because relevance is a fuzzy concept, we used fuzzy logic
methods to represent the semantic distance among concepts
based on each source and then combined the three distances.
Instead of determining if two concepts are related, or what
the chances are that they are related, the system calculates
to what degree or how closely they can be viewed as related.
The calculation of degree of relevance in our system is fre-
quency based: the frequency of occurrence of a relation in var-
ious medical vocabularies, or the frequency of concept
co-occurrence in literature or log data. In the case of medical
vocabularies, consideration is also given to the particular type
of the relation. For instance, ‘‘parent-child’’ relations are con-
sidered to be more important than others.
Once information on related concepts is retrieved from the
three tables, their relevance to the query concept is calculated.
The method ofestimating distance differs slightly byinforma-
A frequency score is assigned to each unique pair of concepts
from a source: Score(Cx, Cy, s), Cx is the query concept, Cy is a
related concepts and s is the source. For relations derived
from MRCOC and QRYLOG, the frequency score of a relation
is simply the frequency of co-occurrence of the 2 concepts in
that relation. For these 2 sources, Score(Cx, Cy, s) 5
Score(Cy, Cx, s).
For relations derived from MRREL, the frequency score is the
weighted co-occurrence of the two concepts in the table.
Because relationships in UMLS are derived from many differ-
ent sources, two concepts can appear as a pair several differ-
ent times. The pair ‘‘heart attack’’ and ‘‘ischemic heart
disease’’ appears 14 times, for instance, under four different
relationships (parent, sibling, broader, and other). Because
we consider the child relationship especially relevant, a
weight of two is given each time a concept is identified as a
child of the query concept. When a parent-child relationship
is involved, Score(Cx, Cy, s)! 5 Score(Cy, Cx, s).
Then, for each related concept a fuzzy score (0 to 1) is com-
puted, representing the degree of relatedness between that
concept and the initial concept. The fuzzy membership, i,
for each set of concepts from a source is defined as:
isðCx;Cy;sÞ 5 0if ScoreðCx;Cy;sÞ 5 0
Cn is any concept that is found to be related to Cx based on
one of the sources. The log transformation is a common tech-
nique used to normalize highly skewed data; we found the
distribution of frequency scores to be highly skewed.
Using A, B, and C to represent the three sources, we com-
bined the degrees of relevance following two fuzzy rules:
1. If two concepts arerelevant in A and B and C, then they are
relevant. (Rule 1)
2. If two concepts are relevant in A or B or C, then they are
relevant. (Rule 2)
For Rule 1, fuzzy intersection of the three fuzzy sets is com-
puted. For Rule 2, fuzzy union of the three fuzzy sets is
The traditional definition of the fuzzy union has been the
maximum function, and the traditional definition of the fuzzy
intersection has been the minimum function. These functions
yield rather crisp results, and when more than two fuzzy sets
are involved they fail to take into account those membership
values between the maximum and minimum.9So we have
used the smoother algebraic sum and algebraic product func-
tions to compute the fuzzy union and fuzzy intersection,
respectively. The membership of an element i in the intersec-
tion of three fuzzy sets, A, B, and C, is defined as the product
of i’s degree of membership in A and i’s degree of member-
ship in B and i’s degree of membership in C:
Fuzzy Intersection 5 iA\B\C5 iA*iB*iC
The fuzzy union is accordingly defined as the algebraic sum
(i.e., the simple sum minus the algebraic products):
Fuzzy Union 5 iA[B[C5 iA1iB1iC2ðiA*iBÞ2ðiA*iCÞ
When translating the membership value into semantic dis-
tance, intersection is given more weight:
Semantic Distance 5 ðiA\B\C31000Þ1iA[B[C
The top n concepts with the shortest semantic distance from a
query concept are considered related to it.
To providean example of the calculation of semantic distance,
Table 2 shows the top 10 related concepts for ‘‘shingles’’ from
each source. Table 3 (Table 3 is available as a JAMIA online
supplement at www.jamia.org) shows the top 10 concepts
Table 1 j Concept Exclusion Criteria for the Different
Source Elimination Rule
MRREL Relationship type 5 ‘‘can be
Relationship type 5 ‘‘is an allowed
Number of co-occurrence , 3
Co-occurrence type 5 ‘‘qualification’’*
Number of co-occurrence , 3
Related concept name . 35 characters
Related concept is a ‘‘stop concept’’y
*The qualification relationship appears to be both overly broad and
yWe maintained a list of ‘‘stop concepts’’ that we have found to be
unusually unhelpful as query concepts. Examples from this ‘‘stop
concept’’ list include the concepts with the names ‘‘Preposition
For,’’ ‘‘Of,’’ and ‘‘With.’’
Journal of the American Medical Informatics Association Volume 13 Number 1Jan / Feb 2006
with the shortest semantic distances to ‘‘shingles’’ and
how they were calculated to take into consideration informa-
tion from each source. The list of query suggestions (i.e.,
Varicella, Herpes Zoster Ophthalmicus, Pneumonia, Preg-
nancy, Neuralgia, Chicken Pox, Ramsey-Hunt Syndrome,
Herpes Simplex, Antiviral, Varicella Encephalitis) that would
be displayed to the user is the list from Table 3. They are
ordered according to the final score they achieved after
computing their scores from each of the three sources.
Concepts that either appear in all three lists or have an ex-
tremely high score in just one of the lists are likely to make
it into the final list.
Semantic-Type Based Recommendations
ications, are common subjects of consumer queries. We found
in a certain aspect of the topics of interest to them, but are not
may be interested in the risk factors for a disease but another
may be interested in the prognosis. To encourage consumers
the system first classifies the concepts based on their semantic
types. For a few major semantic types (e.g., disease and proce-
dure), we identified type-specific modifiers based on pub-
lished literature of consumer HIR needs.1,2The system
suggests some of these modifiers, which have been hard-
coded into the system, if a concept of 1 of these few types
appears in a query. For instance, based on the semantic type
‘‘disease,’’ the system will suggest the concepts ‘‘Symptoms,’’
‘‘Risk Factors,’’ ‘‘Causes,’’ ‘‘Outlook,’’ ‘‘Diagnosis,’’ ‘‘Treat-
ment,’’ and ‘‘Morbidity.’’ If the semantic type is ‘‘procedure,’’
Rate,’’ ‘‘Preparation,’’ ‘‘Indications,’’ ‘‘Complications,’’ and
‘‘Convalescence.’’ These are suggestions based not on the
concept the user entered, but on the type of concept the user
Learning from User Selection
The related concepts identified through fuzzy rules are only
an informed guess of what consumers may find useful in con-
structing a query. The relevance and value of a recommenda-
tion will ultimately be confirmed by usage, which provides a
means for us to improve the quality of the recommendations.
Our system learns from usage by continuously updating the
QRYLOG table: when a suggested concept is selected by a
user, its occurrence with the query concept is increased by
one. The original QRYLOG table only contains concepts
that consumers can recall; the new co-occurrence indicates
what can be recognized. Assuming that ‘‘psoriasis’’ and ‘‘ec-
zema’’ are both in the suggestion list for a query of ‘‘skin’’
with similar ranking, if users consistently click on ‘‘psoriasis’’
but never ‘‘eczema,’’ the co-occurrence of ‘‘psoriasis’’ and
‘‘skin’’ will become higher over time, which in turn will boost
its ranking over ‘‘eczema.’’
HIQuA is implemented using a 3-tier client-server architec-
ture. The client is a Java applet that runs in a Web page, the
middle tier is an Apache Tomcat Web server, and the back
end is a MySQL database server containing millions of med-
ical concepts and relations derived from the Unified Medical
Language System (UMLS), which is provided by the National
Library of Medicine, and the query log data. Queries are sub-
mitted to 1 of several major search engines, with Google?
being the default.
For a given query, HIQuA suggests a list of modifiers and re-
lated concepts. Users may look up definitions of the sugges-
tions, add them to the query, exclude them from the query,
ified query is then submitted to the search engine for free-text
search. Users may also further explore the related concepts of
any recommended concept; these further related concepts are
identified by HIQuA in the same fashion as for the original
query concept. A screen shot of HIQuA is shown in Figure 2.
Usability tests were performed to ensure that the user inter-
face was clearly understood by consumers. We recruited 25
consumers from the Brigham and Women’s Hospital in
Boston to test the system, and iteratively improved the sys-
tem according their feedback. Users sometimes discovered
bugs due to using the system in unexpected ways. They
also pointed out what they found confusing and made
some useful suggestions regarding features they would like
Table 2 j Top 10 Concepts Related to ‘‘Shingles’’ from 3 Sources
Query Term 5 Shingles
Weight of the semantic relationship
with the top 10 concepts from the
medical vocabularies (Methathesaurus
Relationship table, MRREL)
Frequency of co-occurrence with the
top 10 concepts from medical literature
Frequency of co-occurrence
of the top 10 concepts
derived from query log
data (QRYLOG table)
Herpes Zoster Ophthalmicus
Herpes Zoster with Meningitis
Zoster without Complications
Herpes Zoster Iridocyclitis
Skin Diseases, Viral
There are many more results in each of these lists, but only the top 10 are shown here for brevity’s sake.
ZENG ET AL., HIR Query Recommendations
to see, one of which was a navigable history list of searches
We performed a formal evaluation of the system (BWH IRB
Protocol #2003P000710). Consumers who were not health
care professionals were recruited from the Bunker Hill
Community College (BHCC). The eligibility criteria for the
study were some experience with the Web, age 18 or above,
not a physician or nurse, and able to read and write
English. Flyers and posters advertising the study were posted
in the BHCC lobbies. A BHCC computer laboratory room
was borrowed to be the study site. Two discount movie
tickets (approximate $10 value) were given to each subject
who completed the study, which usually required 20 to 30
minutes. The recruitment and testing took place during
June and July of 2004.
A total of 213 subjects participated in the study. All subjects
were asked to use the HIQuA system (in conjunction with
Google?) to search the Web for health information. Query
recommendations were blocked out for half of the subjects
by randomization. Each study subject was asked to first fill
out a brief demographic questionnaire and then perform
1 of 2 predefined health-retrieval tasks—finding five factors
that increase one’s chance of having a heart attack or finding
three methods to treat baldness. We used two different ques-
tions for the predefined task and randomized half of the sub-
jects to each question because it would reduce the chance of
many subjects unexpectedly having prior knowledge of the
given question. The task was described to the participants
Version A: Please find five things that increase a person’s chances
of having a heart attack.
Version B: Some people are concerned about going bald. Please
list three ways to potentially treat their condition.
Each study subject was also asked to perform a self-defined
health-retrieval task. For the self-defined task, subjects were
given verbal instruction to elaborate on their information
needs and retrieval goals in writing prior to the search, so
that we could later evaluate their queries in the context of
the goals. We did not ask for the search results of the self-
defined task to be written down due to practical time con-
cerns (our system did, however, record theirqueries, allowing
us to reconstruct their results later). The self-defined task is
described to the participants as follows:
Please search for any health-related question that you are curious
The subjects were also asked to rate their own overall satisfac-
tion of the search experience on a scale of 1 to 5 at the end of
Please rank your satisfaction with your search experience (circle
1 5 Extremely dissatisfied
2 5 Not satisfied
3 5 Neutral
F i g u r e 2 .
query on the word ‘‘skin.’’ (2) HIQuA displayed a list of query suggestions. (3) The query was sent to Google?. (4) The user
recognized ‘‘psoriasis’’ as the condition he was interested in, so he clicked on that suggestion. (5) HIQuA displayed suggestions
related to psoriasis. (6) Google? displayed search results related to psoriasis.
Screen shots of HIQuA based on the actual search behavior of one of the study subjects. (1) The user submitted a
Journal of the American Medical Informatics AssociationVolume 13Number 1Jan / Feb 2006
4 5 Satisfied
5 5 Extremely satisfied
All queries typed by study subjects and the recommendations
selected by the recommendation group were automatically
recorded in a log file.
To evaluate the impact of HIQuA recommendations, we
measured and compared three main outcomes in the
recommendation group and the nonrecommendation group:
(1) User satisfaction; (2) Query success rate; (3) Score of the
Univariate analysis was first carried out to look at the unad-
justed association between the groups (recommendation vs.
nonrecommendation) and potential demographic factors
including age, race, sex, years of Internet experience, health-
related Web experience, and health status. Only health-
related Web experience and health status were found to be
statistically significant. These two confounders were later
used in the multivariable logistic regression models and the
general linear model to obtain the effect of query recommen-
dations on the three outcomes.
To analyze the first outcome, user satisfaction, the 5-point
user satisfaction ratings were categorized into two categories:
satisfied (including extremely satisfied and satisfied) and not
satisfied (including extremely unsatisfied, unsatisfied, and
neutral). A multivariable logistic regression model was set
up to look at the effect of being in the group receiving query
recommendations on user satisfaction, while controlling for
the confounders. The odds ratios and 95% confidence inter-
vals were computed.
Query Success Rate
To analyze the second outcome, query success rate, a query
that resulted in one or more relevant documents in the top
10 search results was considered successful. A multivariable
logistic regression model was set up to look at the effect of be-
ing in the group receiving query recommendations on the
percent of successful queries, while controlling for the con-
founders. The odds ratios and 95% confidence intervals
In this analysis, we only considered the top 10 results because
too many results were generated by queries for us to assess
recall and precision more comprehensively, and, in any
case, consumers typically only examine the top few search re-
sults.9The success of the queries was evaluated by three
human reviewers: each query (including recommendations
that were clicked on) was submitted to the search engine
and the top 10 returned pages were examined for relevance
based on the pre- or self-defined retrieval goal.
We were able to assess the results for self-defined retrieval
goals because most participants followed the study instruc-
tion and wrote down clear descriptions of their information
needs (and we were able to reconstruct their search results
from the queries recorded in HIQuA’s logs). Every query
and search result was examined by at least two reviewers
and differences between reviewers were resolved through
discussion. For a page to be judged relevant, it needed to
contain at least some information that met the search goal
stated by the participant, and the information could not be
misleading or in the form of commercial advertisement. For
example, for a subject’s question ‘‘How can I prevent sexually
transmitted diseases?,’’ a page on sexually transmitted
disease treatments, a page denouncing abstinence as a
government conspiracy, and a page advertising a particular
brand of condom were all judged to be irrelevant. A total of
280 self-defined tasks were analyzed.
Score of the Predefined Task
To analyze the third outcome, the scoreof the predefined task,
the answers given by subjects for the predefined retrieval task
were graded according to a gold standard that was estab-
lished based on literature review. When grading, a correct an-
swer was given a score of 1, incorrect answers were given a
score of 21, and the absence of an answer was graded as 0.
Incorrect answers were graded lower than the absence of an
answer because being misinformed can be more harmful
than being uninformed. Because we asked subjects to find
and report 5 risk factors for heart disease or 3 treatments
for baldness, all answers to a question were summed up
and divided by 5 or 3, respectively, to generate a normalized
score. Analysis of variance (General Linear Model) was used
to compare the predefined task score of the group receiving
query recommendations versus the group that did not receive
query recommendations. We adjusted for the 2 confounders
in the analysis.
A total of 213 subjects participated in the study. We had a
fairly diverse population of subjects in terms of race and eth-
nicity. On average, the subjects were young, reasonably well
educated, and healthy (Table 4 is available as a JAMIA online
supplement at www.jamia.org). Please note that the education
level indicates the highest level started, not finished. Many of
the subjects were attending the community college where we
conducted our study. Over 40% did not speak English as
their first language, andthe commandof Englishvaried signifi-
cantly among these non-native speakers. The subjects were
generally familiar with the Web, though not all had had Web
Of those in the group receiving query recommendations,
85.2% of the subjects were satisfied with their search experi-
ence—a result that was a little higher than for the nonrecom-
mendation group (80.6%). However, the difference was not
statistically significant (p 5 0.136). According to the odds ra-
tio calculated using logistic regression, the odds of being
satisfied increased by 79% if the participant was in the
recommendation group. The confidence interval for the
odds ratio, however, is wide and crosses 1.0; thus the associ-
ation between groups and user satisfaction is not statistically
significant (Table 5).
Query Success Rate
There was a statistically significant difference (p 5 0.006) in
the percentage of successful queries between the recommen-
dation group (76.0%) and the nonrecommendation group
(65.7%) (Table 5). According to the odds ratio calculated using
logistic regression, being in the recommendation group
increased the odds of submitting a successful query by
66% (with a 95% confidence interval of between 16% and
ZENG ET AL., HIR Query Recommendations
Because a statistically significant difference was found for the
query success rate between the recommendation group ver-
sus the nonrecommendation group, we further examined
the source of the difference. The queries manually typed in
by both groups of subjects did not have a statistical difference
in success rate, as one would expect. The suggested queries
that were selected by subjects in the recommendation group
did have a higher success rate (p , 0.0001) than the typed-
in queries (Figure3 is available as a JAMIA online supplement
at www.jamia.org). (This comparison was also adjusted for
the two confounders—health-related Web experience and
Score of Predefined Task
The normalized mean score of the predefined task was higher
for the nonrecommendation group (0.577) than for the recom-
mendation group (0.440), although not statistically significant
(p 5 0.233). In Table 5, we report both mean and median for
the third outcome because the distribution of scores was
To summarize, the use of query recommendations led to a
higher rate of successful queries. The impact (positive or neg-
ative) of query recommendations on satisfaction or accom-
plishing a predefined retrieval task was not clear.
There are known limitations to our development and evalua-
tion methodology. First, the target users are consumers,
which is a very diverse group. It can be argued that each con-
sumer has a different mental model; however, even a diverse
group shares common terms, concepts, and concept relations.
Take the term ‘‘anorexia,’’ for instance—in the professional
setting it usually refers to the symptom ‘‘loss of appetite’’
while in the lay setting it usually refers to the disease anorexia
nervosa. We use the adjective usually here because there are
always exceptions. Yet if there did not exist some common
mental model among consumers, and between consumers
and physicians, it would be impossible for physicians to com-
municate with consumers and for consumers to communicate
with each other. Nevertheless, the diversity of the consumer
population makes measurement of semantic distance be-
tween concepts inherently less precise.
For query expansion, consumer queries are mapped to UMLS
concepts by string matching. Accurate mapping is not always
feasible because the UMLS concepts and concept names pri-
marily represent the language of health care professionals.
We are currently involved in a collaborative effort (www.
consumerhealthvocab.org) to address this very issue. Disam-
biguation also remains challenging. We may disambiguate
incorrectly and thus present the user with unhelpful
suggestions. In these cases, the user is free to ignore the sug-
gestions. Note that without being privy to the internal
thoughts of the user, we often cannot know whether we
have disambiguated incorrectly, e.g., perhaps a particular
user actually desires to find information on a genus of crabs
when entering the query ‘‘cancer.’’
Our and others’ previous analyses of consumer health queries
and online postings showed that about 50% to 80% of con-
sumer query terms can be mapped to UMLS.5,42,43Although
this mapping rate is not ideal, it provides a starting point for
our concept-based query expansion. In our identification of
related concepts, three sources are involved which have a
small common overlap (less than 5% by our observation)
while being largely complementary to each other. Having a
related concept declared relevant by more than one source
or having one source rank a related concept extremely high
suggests a shorter semantic distance. The rules we used are
a fuzzy representation of this basic logic, which is not equiv-
alent to an algebraic mean of rankings from each source. We
acknowledge that this might not be the optimal solution, but
rather, a solution which reflects the intuitive ways in which
people combine information from multiple sources. (There
is no universalsolutionto
combining semantic-distance information from multiple
sources—different approaches apply to different domains.)
the generalproblem of
Regarding the second outcome measurement (query success
rate), the query success was determined by the researchers in-
stead of study participants. Researchers judged the relevance
of a page based on whether it met the retrieval goals stated by
participants, or the predefined retrieval goal. A potential
problem of this approach is that researchers could make mis-
takes in interpreting the retrieval goals written by partici-
pants, although most goals were relatively straightforward,
e.g., ‘‘How can I prevent sexually transmitted diseases?’’ On
the other hand, researchers tend to be more consistent and
better equipped to review the relevance of a page of health in-
formation than study participants.
Time spent by participants conducting the searches was re-
corded, but not reported as an outcome. One reason is that
we found that there could be different causes for spending
more time at a task: it sometimes resulted from finding inter-
esting material to read and explore and sometimes from not
being able to find the desired information.
Finally, we did not distinguish officially published literature
from unpublished literature (‘‘grey literature’’) in this study,
and neither did we distinguish high-quality from low-
quality material. The quality and credibility of content are
important issues, but it was beyond the scope of HIQuA
Table 5 j Comparison of User Satisfaction, Query Success, Predefined Task Score of the Recommendation Group
and Nonrecommendation Group
Group p Value
and Odds Ratio*
Query success rate
Predefined task scorey
20.137 (20.320 to 0.049)0.440 6 0.702
0.577 6 0.653
*95% confidence interval for estimated odds ratios and mean difference from logistic regression and linear model.
yMean 6 standard deviation, and median.
Journal of the American Medical Informatics AssociationVolume 13Number 1 Jan / Feb 2006
We have designed, implemented, and evaluated a tool to help
consumers with query construction duringHIR. The resulting
specific or complex queries. The HIQuA system uses fuzzy
logic to combine semantic distance information from three
sources (concept co-occurrence in query log and medical
literature, and semantic relationships in medical vocabular-
ies), for the purpose of identifying relevant concepts. It also
learns from user selection to continuously refine the recom-
mendations. The evaluation showed that the availability of
recommendations led to a significantly higher rate of success-
ful queries, although there was not any significant impact on
user satisfaction or on accomplishing a predefined retrieval
Because HIQuA can be used to explore the semantic neigh-
borhood of tens of thousands of medical concepts, consumers
may first browse the concept space to find the right term(s) to
describe their needs and then look in the content space for the
relevant information. There exist Web directories that con-
sumers can browse for health information, but these directo-
ries mostly reflect hierarchical or classification knowledge
regarding medical concepts. HIQuA constructs a concept
neighborhood based on a much broader scope of medical
knowledge and takes consumer usage patterns and consumer
mental models into account.
In presenting the related terms to users, we did not simply
use the UMLS preferred name because many preferred names
are not the most user-friendly among all the synonyms. We
have identified a set of consumer-preferred names for tens
of thousands ofUMLS concepts primarily basedon how often
a name is used by lay people. These names are used whenever
available as the display names for concepts. They are also
naturally free of the ‘‘NOS’’-type postfixes present in some
vocabularies, because no consumer ever adds a ‘‘,1.’’ or
‘‘NOS’’ behind a term. (‘‘NOS’’ stands for not otherwise
specified; ‘‘,1.’’ is sometimes added by a vocabulary to in-
dicate that a certain string is preferred for one concept over
Without knowing the context of a query, HIQuA makes rec-
ommendations on the basis of two postulations: (1) a user
may want to refine or replace the search term with other
related terms; (2) the relatedness of terms can be derived
from co-occurrences in usage and from known semantic rela-
tions. HIQuA is limited in its capacity to understand the real
information needs underlying a query, especially a short one.
It thus can only make best guesses about which other terms
might be of use to a consumer.
The evaluation showed that HIQuA recommendations
helped consumers to generate more successful queries, which
helped to validate the design and implementation of the sys-
tem. Several factors contributed to our failureto show a statis-
tically significant impact of the system on overall user
satisfaction or on the score of the predefined task. First, not
every consumer needs the help of recommendations when
performing every single task. Some subjects in the nonrecom-
mendation group can accomplish the given or self-defined
retrieval tasks successfully on their own. Second, not all sub-
jects made use of the recommendations. Six people in the
recommendation group did not click on any recommenda-
at least one subject clicked on every query term suggested by
HIQuA, many of which did not help with the retrieval tasks.
Third, query recommendations would not be of help to peo-
ple with very poor health literacy and very poor general
literacy levels. Several study subjects misinterpreted the pre-
defined question or the information they had found: a few
subjects wrote down causes for baldness although the ques-
tion was how to treat the condition. Some subjects clearly
were unable to discern the promotional or misleading infor-
mation from ‘‘good’’ information and thus gave wrong an-
swers. Fourth, satisfaction is a very subjective measurement.
Some people answered the predefined question completely
incorrectly, yet reported satisfaction with the search experi-
ence. Because of these factors and the sample size, it was un-
derstandable that a statistically significant difference on the
user satisfaction score or on the predefined task score was
not found between the recommendation and nonrecommen-
dation groups. A larger sample size might have resulted in
statistically significant findings.
The innovation of the HIQuA system is that it estimates se-
mantic distance based on three types of information sources
(i.e., query log, literature corpus, and manually constructed
science research has explored each of these types individually
section, there have also been studies that used multiple infor-
mation sources and utilized fuzzy logic in query expansion.
tiple co-occurrence data with relations from vocabularies.
In the specific area of consumer HIR, research on query ex-
pansion or suggestion has depended on medical vocabularies
as the main knowledge source. However, the HIQuA system
explores other sources and makes use of semantic relations
beyond synonymy and hierarchical relationships. The use of
the query log is especially important because it is a record
of consumer language and consumer search behaviors.
Query suggestions will not be needed by every user for every
search; however, the evaluation has shown that our system
could be a helpful tool for query formation when a user
does need it. Because there are millions of consumers con-
ducting HIR, even a fraction of the entire user population
comprises a great number of users. As a general purpose ap-
plication in the health care domain, HIQuA could potentially
benefit many users conducting HIR.
To help consumers obtain better satisfaction and retrieval per-
formance when querying, we will continue to work on the re-
finement of this system as well as some other HIR issues such
as content annotation and quality assessment.
We have developed a query suggestion tool to help con-
sumers search for health information online. Our approach
is designed to address the problems of user query construc-
tion by providing frequency- and knowledge-based query
recommendations. Our trial showed that providing HIQuA
recommendations resulted in statistically significantly more
successful consumer queries over not providing the recom-
mendations, although no statistically significant impact on
user satisfaction or ability to accomplish a predefined
ZENG ET AL., HIR Query Recommendations
retrieval task was found. Although query expansion has been
studied extensively, using fuzzy logic to combine information
derived from usage logs, literature co-occurrence, and vocab-
ulary information is novel. While prior research in query ex-
pansion or query recommendations for consumer HIR has
been mainly thesaurus-based, our study tested the feasibility
of (and showed promising results for) employing more di-
verse sources to find related terms or concepts. Because query
formation is a challenging task for many HIR users, we be-
lieve that our system, or a similar system, could have a posi-
tive impact on HIR for consumers by providing meaningful
and consumer-centered suggestions.
1. Fox S, Fallows D. Health searches and email have become more
commonplace, but there is room for improvement in searches
and overall Internet access: Pew Internet & American Life Proj-
ect; 16 July 2003.
2. Zeng QT, Kogan S, Plovnick RM, Crowell J, Lacroix EM, Greenes
RA. Positive attitudes and failed queries: an exploration of the
conundrums of consumer health information retrieval. Int J
Med Inform. 2004;73:45–55.
3. Silverstein C, Marais H, Henzinger M, Moricz M. Analysis of a very
4. Spink A, Wolfram D, Jansen MBJ, Saracevic T. Searching the
Web: the public and their queries. J Am Soc Inform Sci Technol.
5. Zeng Q, Kogan S, Ash N, Greenes RA, Boxwala AA. Character-
istics of consumer terminology for health information retrieval.
Methods Inf Med. 2002;41:289–98.
6. Kogan S, Zeng Q, Ash N, Greenes RA. Problems and challenges
in patient information retrieval: a descriptive study. Proc AMIA
7. Eysenbach G, Kohler C. How do consumers search for and ap-
praise health information on the World Wide Web? Qualitative
study using focus groups, usability tests, and in-depth inter-
views. BMJ. 2002;324:573–7.
8. Fredin ES, David P. Browsing and the hypermedia interaction
cycle: a model of self-efficacy and goal dynamics. Journal
Mass Commun Q. 1998;75:35–55.
9. Baeza-Yates R, Ribeiro-Neto B. Modern information retrieval.
New York, NY: ACM Press; 1999.
10. Cui H, Wen J-R, Nie J-Y, Ma W-Y. Query expansion by mining
user logs. IEEE Trans Knowledge Data Eng. 2003;829–39.
11. Stenmark D. Query expansion on a corporate intranet: using LSI
to increase precision in explorative search. In: Proceedings of the
38th Annual Hawaii International Conference on System Sci-
ences; 03-06 Jan. 2005; Page 101c.
12. Harman D. Towards interactive query expansion. In: 11th An-
nual International ACM SIGIR Conference on Research and De-
velopment in Information Retrieval, May 1988; Grenoble,
France: ACM Press, 1988, p. 321–31.
13. Xu J, Croft WB. Query expansion using local and global docu-
ment analysis. In: The Nineteenth Annual International ACM
SIGIR Conference on Research and Development in Information
14. Mandala R, Tokunaga T, Tanaka H. Combining multiple evi-
dence from different types of thesaurus for query expansion.
In: The 22nd Annual International ACM SIGIR Conference on
Research and development in information retrieval; 1999; Berke-
ley, CA, 1999:191–7.
15. Lee H-M, Lin S-K, Huang C-W. Interactive query expansion
based on fuzzy association thesaurus for Web information re-
trieval. In: Fuzzy systems, 2001. The 10th IEEE International
Conference, 2001:724–7, vol 3.
16. Takagi T, Tajima M. Query expansion using conceptual fuzzy
sets for search engine. In: Fuzzy systems, 2001. The 10th IEEE
International Conference, 2001:1303-8.
17. Akrivas G, Wallace M, Andreou G, Stamou G, Kollias S. Con-
text-sensitive semantic query expansion. In: Artificial Intelli-
gence Systems, 2002. (ICAIS 2002). 2002 IEEE International
18. Anick P. Using terminological feedback for web search refine-
ment: a log-based study. In: The 26th Annual International
ACM SIGIR Conference on Research and Development in Infor-
mation Retrieval, Toronto, Canada, 2003:88–95.
19. Muramatsu J, Pratt W. Transparent queries: investigating users’
mental models of search engines. In: The 24th Annual Interna-
tional ACM SIGIR Conference on Research and Development
in Information Retrieval, New Orleans, LA, 2001:217–24.
20. Ruthven I. Re-examining the potential effectiveness of interac-
tive query expansion. In: The 26th Annual International ACM
SIGIR Conference on Research and Development in Information
Retrieval, Toronto, Canada, 2003:213–20.
21. Billerbeck B, Scholer F, Williams HE, Zobel J. Query expansion
using associated queries. In: The Twelfth International Confer-
ence on Information and Knowledge Management. New
Orleans, LA: ACM Press, 2003. pp. 2–9.
22. Doszkocs TE. AID, an associative interactive dictionary for
online searching. Online Rev. 1978;2:163–72.
23. Pollitt S. CANSEARCH: an expert systems approach to docu-
ment retrieval. Inform Process Manage. 1987;23:119–38.
24. Srinivasan P. Retrieval feedback in MEDLINE. J Am Med Inform
25. Hersh W, Price S, Donohoe L. Assessing thesaurus-based query
expansion using the UMLS Metathesaurus. Proc AMIA Symp.
26. Aronson AR, Rindflesch TC. Query expansion using the UMLS
Metathesaurus. Proc AMIA Annu Fall Symp. 1997;485–9.
27. Nenadic G, Mima H, Spasic I, Ananiadou S, Tsujii J. Terminol-
ogy-driven literature mining and knowledge acquisition in bio-
medicine. Int J Med Inform. 2002;67:33–48.
28. Joubert M, Fieschi M, Robert JJ, Volot F, Fieschi D. UMLS-based
conceptual queries to biomedical information databases: an
overview of the project ARIANE. Unified Medical Language
System. J Am Med Inform Assoc. 1998;5:52–61.
29. Johnson SB, Aguirre A, Peng P, Cimino J. Interpreting natural
language queries using the UMLS. Proc Annu Symp Comput
Appl Med Care. 1993;294–8.
30. Merz RB, Cimino C, Barnett GO, Blewett DR, Gnassi JA, Grund-
meier R, et al. Q & A: a query formulation assistant. Proc Annu
Symp Comput Appl Med Care. 1992;498–502.
31. Salton G, Buckley C, Fox EA. Automatic query formulations in
information retrieval. J Am Soc Inf Sci. 1983;34:262–80.
32. Hripcsak G, Allen B, Cimino JJ, Lee R. Access to data: comparing
AccessMed with Query by Review. J Am Med Inform Assoc.
33. Fisk JM, Mutalik P, Levin FW, Erdos J, Taylor C, Nadkarni P.
Integrating query of relational and textual data in clinical data-
bases: a case study. J Am Med Inform Assoc. 2003;10:21–38.
34. Gobel G, Andreatta S, Masser J, Pfeiffer KP. A MeSH based intel-
ligent search intermediary for Consumer Health Information
Systems. Int J Med Inform. 2001;64:241–51.
35. McCray AT, Ide NC, Loane RR, Tse T. Strategies for supporting
consumer health information seeking. Medinfo. 2004;11(Pt 2):
36. Plovnick RM, Zeng QT. Reformulation of consumer health
queries with professional terminology: a pilot study. J Med Inter-
net Res. 2004;6(3):e27.
37. Patrick TB, Monga HK, Sievert ME, Houston Hall J, Longo DR.
Evaluation of controlled vocabulary resources for development
of a consumer entry vocabulary for diabetes. J Med Internet
Journal of the American Medical Informatics AssociationVolume 13 Number 1 Jan / Feb 2006
38. Mendonca EA, Cimino JJ. Automated knowledge extraction
from MEDLINE citations. Proc AMIA Symp. 2000;575–9.
39. Hristovski D, Stare J, Peterlin B, Dzeroski S. Supporting discov-
ery in medicine by association rule mining in Medline and
UMLS. Medinfo. 2001;10(Pt 2):1344–8.
40. Humphreys BL, Lindberg DA, Schoolman HM, Barnett GO. The
Unified Medical Language System: an informatics research col-
laboration. J Am Med Inform Assoc. 1998;5:1–11.
41. Miller N, Lacroix EM, Backus JE. MEDLINEplus: building and
maintaining the National Library of Medicine’s consumer health
Web service. Bull Med Libr Asnsoc. 2000;88:11–7.
42. Tse T. Identifying and characterizing a ‘‘consumer medical vo-
cabulary.’’ College Park, MD: University of Maryland, 2003.
43. Brennan PF, Aronson AR. Towards linking patients and clinical
information: detecting UMLS concepts in e-mail. J Biomed In-
ZENG ET AL., HIR Query Recommendations