Electronic Records for General Practice

Where we Are, Where we should Head to Improve Them

Federico Cabitza

, Francesco Del Zotti

and Paolo Misericordia

Dipartimento di Informatica (DISCo), Università degli Studi di Milano-Bicocca, V.le Sarca 336, Milano, Italy

NetAudit, Verona, Italy

Centro Studi FIMMG, Piazza Guglielmo Marconi 25, Roma, Italy

Keywords: Electronic Medical Records, General Practice, Quality Evaluation, Online Questionnaire, Attitude Survey.

Abstract: In this paper we present the findings of a country-wide survey that was aimed at getting a comprehensive

picture of the current level of adoption and appropriation of Electronic Medical Records (EMRs) by General

Practitioners (GPs). In this survey, which collected the responses from 800 Italian GPs coming from all over

the country and exhibiting different experience and ICT skills, we investigated the level of current

satisfaction of users with respect to two classes of functional and non-functional features of current EMRs,

namely “core” and “advanced” ones. We also tried to detect which of these features are valued most highly

by the current EMR users in order to inform prospective users, EMR vendors and policy makers in the

eHealth domain. We also focused on the impact that digitization has had so far on General Practice, as it is

perceived from the perspective of the front-line users (i.e., doctors). Finally we also addressed how the use

of ICT could change in the near future as a tool to facilitate doctor patient communication and collaboration.

1 MOTIVATIONS AND

BACKGROUND

General Practice is one of the professions that in the

last years have been affected more deeply by a

process of continuous digitization of its document-

and communication-related tasks (Benson, 2002;

Delaney, 2010; Dobrev et al., 2008; Purves, 1996).

The Electronic Medical Record (EMR) is the main

tool at the centre of this process of digitization, and

therefore it has so far received the interest of

hundreds of scholarly studies aimed at

understanding the nature of its role in changing

medical work, e.g., (Hassey, 2001; Hippisley-Cox,

2003; Porcheret et al., 2004; Treweek, 2003). For

this reason, we also aim our research to get a general

picture of how General Practitioners (GPs) perceive

their electronic records in 2013, that is in an age

when new widespread portable devices have rebuilt

the concept of mobility, the Web 2.0, and the Social

Media it helps spreading, have changed expectations

in patients and users in general, and data exchange

and interoperability have finally come of age,

showing unprecedented results and reliability. In this

context of new technologies in place and “old” ones

that have reached full maturity and convinced even

the so called “late majority” of potential users, we

want to assess: their level of adoption (i.e., use); the

perceived appropriation by their users (e.g., how

they fit their needs), and the satisfaction of the GPs.

In what follows we will then report on an attitude

user study that involved a vast sample of primary

care doctors and probed them on: satisfaction and

usefulness of both traditional and more advanced

aspects of their digitial media; the impact they

believe digitization has had on their medical

practice; what areas of improvement ICT designers

and developers should focus on to provide them with

better tools; the functionalities that they perceive as

most useful; and lastly, how they currently use ICT

to communicate and collaborate with their patients

and how they intend to use it in the near future.

We believe that iterating this kind of initiatives on a

periodic basis could allow ICT designers and policy

makers to better fit the real needs of the end users

involved and hence to affect quality of care, user

satisfaction and ICT appropriation positively.

2 METHODS

To address the research question mentioned above,

535

Cabitza F., Del Zotti F. and Misericordia P..

Electronic Records for General Practice - Where we Are, Where we should Head to Improve Them.

DOI: 10.5220/0004924805350542

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2014), pages 535-542

ISBN: 978-989-758-010-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

we focused on three main dimensions of analysis:

1. the impact of EMRs on the work of the GPs,

2. the quality perceived by GPs in regard to their

EMR, and

3. the use of communication technology by GPs.

The perceived impact of EMRs on general

practice has been articulated in relation to typical

work dimensions, like productivity,

knowledgeability, collaboration, stress and

appropriateness; and also in terms of the extent ICT

supports typical phases (or “concerns”) of medical

work, like “prevention”, “examination”, “diagnosis”,

“treatment”, “follow-up” and “audit/evaluation”.

These latter items have been grouped in a specific

construct called “QSM”. The perceived quality of

EMRs has been analysed in regard to three aspects:

i) overall recommendability; ii) satisfaction, and iii)

perceived utility. Recommendability is considered in

terms of Net Promoter Score (Keiningham et al.,

2007). Satisfaction has been further articulated in

terms of two constructs (set of items): QSC, QSA.

The construct called “QSC” regards perceived

satisfaction for those aspects (i.e., functions and

attributes) that EMRs usually cover: namely “input

facilitation”, “personalization”, “reliability”,

“performance”, “usability”, “level of integration”,

“adequacy of the information model”.

The construct we call “QSA” regards more

“advanced” functionalities: namely “support of

mobility”, “multiple sharing capabilities”,

“automatic import”, “support of semi-automatic

audit”, “flexible back-up management”, “case- and

process-aware reminders”, “knowledge source

search and retrieval”, “full record annotation”.

The perceived utility of the “advanced” functions

of EMRs has been expressed by a set of items

encompassed by the “QUA” construct, whereas the

utility of the core functionalities were given for

granted. Finally, the role of ICT in mediating

communication and collaboration among GPs and

their patients is articulated in terms of both current

use and desired prospective use, in order to see how

ICT could affect the interaction between GPs and

their patients also in the near future.

In August 2013, we administered a questionnaire

reflecting the information model described above to

a selected sample of Italian GPs, whose addresses

were indexed in the mailing list of the “Federazione

Italiana Medici di Medicina Generale” (FIMMG).

The FIMMG is a trade union organization and

country-wide professional association that represents

almost 30,000 General Practitioners in Italy. With

the collaboration of the Study Center of this

association, we contacted 14,478 potential

respondents by sending them one single message of

invitation by e-mail. In this message we invited the

recipients to participate in the online survey on a

voluntary and anonymous basis, with no incentives

nor reminders, once the main aims of the research

had been outlined. Responses were collected for 5

weeks through an online questionnaire platform,

Limesurvey v. 2.0. We configured this platform to

have it display the questionnaire as a sequence of six

short Web pages that enabled a Computer-Assisted

Self Interview (CASI). In these pages, we kept the

number of mandatory fields to a minimum to only

allow for the conditional routing of items, so that the

overall interview could take a shorter time on the

basis of the responses provided in the process. The

platform also allowed for unique and restricted

access to the questionnaire, so that no multiple

responses were collected, and only invited subjects

could participate in the survey until we closed it.

The platform also enabled data persistence across

multiple user sessions to allow respondents to quit

the questionnaire anytime and resume it at a later

time. In doing so, we aimed to minimize the risk of

fatigue bias: nevertheless, the pilot sessions by

which we tested the survey with a small convenience

panel before the invitation had been dispatched took

between 8 and 12 minutes on average to complete.

3 RESULTS

At the end of the survey, we collected 800 complete

questionnaires. Respondents (77% men and 23%

women) were those who accepted to participate in

the interview mentioned above and filled in all the

items reported in the questionnaire. Partial

questionnaires were then discarded. The relatively

low response rate was also due to a technical

problem that occurred in the dispatch of the

invitation letter, which caused many letters be

erroneously categorized as spam by aggressive

filters, and to the fact that the survey was

administered when many practitioners were likely

on a summer vacation.

The average age of the respondents was 57 years

(sd: 6 yrs, range: 27 – 70 yrs, 95% c.i.: 56.5 – 57.3)

and 86% of the sample declared an age between 50

and 65 years. 91% of respondents claimed to have

more than 10 years of experience in the GP field,

and 73% more than 20 years. 85% of the sample

declared to assist more than 1000 patients (22%

even more than 1500; average number of patients:

1290 ± 340). These data make us confident to have

collected the opinions and attitudes of experts of

HEALTHINF2014-InternationalConferenceonHealthInformatics

536

great experience and deep knowledge of their

application domain.

Italian macro Regions were represented in the

sample with a slight preponderance of Northern

Regions with respect to actual distributions of Italian

GPs (North: 49% vs. 43%, Centre: 20% vs. 21%,

South: 32 vs. 36%). Thus, we weighted our sample

in order to make the sample representative at

geographical level. Margin of error is 3.4% at

National level (at a confidence level of 95%). The

detected margin is consistent with conservative

assumptions on response distribution, since the

research does not claim census-like aims, but rather

it aims to detect attitudes and trends in EMR use and

its perceived quality. This makes our study relevant

to build a picture of current EMR adoption and

satisfaction and to be considered by the scientific

community of IT researchers for the aims mentioned

in Section 1.

In regard to IT use, 98% claimed to be using an

Electronic Medical Record. This is consistent with

other similar recent surveys, e.g., (Cabitza, 2012)

(Misericordia, 2011). However, among the users that

claimed to be using an EMR, the proportion of those

who claimed to be exploiting their record only “for a

small part of its potential” and, anyway, “less than

actually felt necessary” is (still) high (see Figure 1,

topmost diagrams): in either cases, approximately 1

GP out of 4. This is in line with the concerns

discussed in (Cabitza, 2012) regarding the

phenomenon of “low use”(Simon et al., 2009), and

the related efforts to pay to improve user adoption

and appropriation (our model addresses these points

in two specific items, called PerceivedExploitation

and PerceivedFit). The 94% of the sample found

their EMR listed in the questionnaire, which

reported 15 applications from the Italian market.

Three applications were used by nearly the two

thirds of the sample: we will call them Application

A (used by 45% of respondents), B (12%) and C

(11,8%) for sake of confidentiality.

Figure 1 depicts other characteristics of the

respondent sample, in particular with regard to

declared familiarity with IT (Perceived Skills), and

to what kind of cooperative relationships GPs are

involved in their practice (Collaboration Type): this

latter information is relevant with respect to the

granularity of the “sharing” functionality discussed

in the QSA and QUA constructs of the model.

All of the aggregated constructs mentioned

above showed very high internal consistency in

terms of intraclass correlation: satisfaction in regard

to support of medical dimensions (QSM, Cronbah’s

alpha=0.89, c.i.= .88-.90, mean scale=18, sd=4.4);

Figure 1: Descriptive statistics of the respondent sample.

Figure 2: Inter-construct correlations.

satisfaction for common functions (QSC, Cronbah’s

alpha=0.84, c.i.=.82-.85, scale mean=16, sd=3.9);

satisfaction for advanced functions (QSA,

alpha=.92, c.i.=.87-.96, scale mean=40, sd=13);

perceived utility for advanced functions (QUA,

alpha=.98, c.i.=.96-.99, scale mean=49, sd=16),

ElectronicRecordsforGeneralPractice-WhereweAre,WhereweshouldHeadtoImproveThem

537

perceived impact on work dimensions (PIW,

alpha=.79, c.i.=.76-.81, mean=28, sd=4.2).

Canonical correlations between constructs were

significant between QSC and QSA (alpha=.94),

QSC and QUA (=.91), QSM and QSA (.86).

Correlations between QSM and QSC, and between

QSM and QUA are slightly below the conventional

significance threshold (.63 and .72 respectively).

Figure 2 shows these inter-construct correlations in

graphical form. In what follows we report the main

findings for each construct of analysis.

3.1 Recommendability

As anticipated above, in regard to overall

recommendability we adopted the Net promoter

Score (NPS). This is a convenient way to represent

satisfaction for a product as this score relates

satisfaction to the pragmatic behaviour of their users

and associated likelihood to recommend such a

product to colleagues and acquaintances. The overall

NPS of the EMRs was +13%; this means that

promoters, i.e., who would highly recommend her

application responding either “9” or “10” on a 1-to-

10 scale, were 13% more numerous than detractors,

i.e., who would not recommend her application (and

responded with a mark below “6”). Nevertheless,

situation differs according to the application actually

owned. Just to limit ourselves to the three most used

EMRs (A, B and C mentioned above), the NPS of A

was +35% (i.e., the proportion of potential

promoters was one third larger than the proportion

of potential detractors), C’s NPS was +11%, and B’s

one was -30% (i.e., the proportion of potential

detractors was one third larger than the proportion of

potential promoters). Moreover, if we assimilate the

NPS 10-point ordinal scale to an interval scale, we

could evaluate an “overall recommendability” score

of current Italian EMRs in terms of “average score”,

that is 7.6 (± 2). In regard to the most used EMRs,

the average score is 8.2 for A, 6.4 for B, and 7.6 for

C (the difference between the scores of A and B is

highly significant, t=2.9, P=.004). The Net Promoter

Score of the most widely adopted EMR (45%)

resulted to be highly affected by self-perceived IT

skills (P-value<.001, the higher the skills, the higher

the NPS), perceived level of exploitation (P-

value<.001, the higher the level, the higher the NPS)

and, notably, the number of patients (P=.001, the

higher the number of patients, the higher the NPS);

also geographic and collaboration-related

differences could have an impact (respectively,

P=.07 and P=.09) but this is not statistically

significant; other factors, like gender, age and years

of experience result not related to NP scores.

3.2 Satisfaction for Common Features

For sake of clarity, we report the results for the QSC

construct with the aid of Figure 3. There we depict

the response distribution for the question expressed

as follows: “in regard to each feature would you

please indicate the extent your current EMR should

be still improved to really please you”. Available

options ranged on an four item scale, encompassing

the labels “still a lot of room for improvement” (1),

“some room for improvement” (2), “already

adequate” (3), and “already optimal” (4). Boxplots

show the interquartile range for perceived

satisfaction: the dark green box indicates the upper

quartile; the light green box the lower quartile; thus,

the median is depicted as the border line between the

two quartile boxes (if visible); whiskers indicate the

variability outside the upper and lower quartiles (i.e.,

the min and max value respectively). Moreover, as a

merely illustrative addition, we also indicate the

response average with a small red dash, and the

mode (i.e., the response chosen by the majority of

respondents) with a blue dash.

Figure 3: Box plots showing the interquartile range for

QSC.

In the topmost part of Figure 3, we also show the

result of a Binomial test that we undertook to

understand whether the apparent tendency that can

be detected for each item visually is also “true” and

generalizable to the whole population of reference,

or it is not so because, e.g., that tendency is due to

chance. The little hand-shaped icons indicate

whether the respondents were sufficiently satisfied

HEALTHINF2014-InternationalConferenceonHealthInformatics

538

with the corresponding feature (i.e., responses

associated with at least an “adequate” level of

satisfaction) or not (i.e., responses either below or at

the threshold of satisfaction, that is 2). The number

of asterisks indicates how much “significant” the

indication is according to the Binomial test.

According to the results, then, the GPs indicated

clearly that current EMRs should be improved with

respect to:

1) their capability to make simple tasks more

automatic, like offering more precompiled

fields in form templates, the capability to

duplicate data between fields when applicable,

and more user-friendly import and export

functionalities.

2) their capability to be customized by end-users,

both in terms of user interface (appearance,

what fields and where to position them,

available formats) and functional

configurability.

3) their performance, both in terms of speed in

processing data and in reacting to user

commands and interactions.

4) their compatibility with data coming from

different applications and their interoperability

with other systems and services, also at

Regional level.

5) the adequacy of the information model by

which patient cases and illness trajectories are

represented in the electronic record (on this

problem, the interested reader can refer to

(Swinglehurst et al., 2012)).

Respondents also expressed the need to improve

the reliability of their EMR, in terms of either

absence of errors and failures. Quite surprisingly,

respondents claimed they are already satisfied with

the usability of the graphical interface and the

overall user experience, which conversely is a

common place of user dissatisfaction.

It is worthy of note the fact that both the

perceived fit of EMRs with the GPs’ needs and the

perceived level of exploitation seem strongly

correlated with the degree of satisfaction for several

functions. Other contextual aspects seem to play a

less important role as factors influencing

satisfaction, except the number of patients treated by

the single GP: the higher this number, the higher the

satisfaction with respect to usability and the

information model.

3.3 Perception of Advanced Features

Advanced features have been investigated both in

terms of actual satisfaction (QSA) and perceived

usefulness (QUA).

In regard to satisfaction for a specific feature,

this was assessed on an ordinal scale of 6 items,

from 1 (totally unsatisfied) to +6 (totally satisfied).

Quite surprisingly, the GPs involved in our study

expressed a generally higher appreciation for the

advanced features than for the core functionalities

(represented in the previous construct, QSC). It is as

if car drivers showed appreciation for the luxury

accessories of their cars (e.g., the leather-trimmed

interiors), but expressed discontent for the car’s

reliability and its low compatibility with regular

petrol stations. More seriously this could be related

to the high experience of the respondents involved,

probably aware of what can really produce value in

their professions (on which they are demanding in

reason) and what is fine but somehow superfluous.

Like in the case of the QSC construct, we

verified QSA trends performing a Binomial test for

each feature, by comparing the proportions of

responses that were either above or below the

satisfaction “threshold” (+3). Respondents claimed

to be already satisfied in regard to the support of

mobility, the content sharing with colleagues (other

GPs), the capability to execute audits on the basis of

their own patient data, to perform backup, to set

proactive reminders, and to annotate their records.

On the other hand, the data analysis could not detect

a clear satisfaction for the capability to share data

with patients, and to import data from other systems.

A clear unsatisfaction could not be proved

significant with respect to the capability to share

information with colleagues of the “continuous

assistance” program (CA, a sort of "out of hour"

medical services) and colleagues other than GPs

(e.g., specialists, hospital doctors). On the other

hand, significant unsatisfaction has been expressed

for the service by which to retrieve the latest

scientific sources (usually journal articles) from the

specialist literature that are pertinent to specific

patient cases.

As mentioned above, the construct called QUA

regards the perception of usefulness by the sample

and it is expressed in terms of items whose possible

answer options were represented on a six-value

ordinal scale, from 1 “useless at all” to 6 “very

useful”. We intend this construct as an informal but

consensus-based indication of what functionalities

should be addressed first by EMR vendors, and

therefore should be added to the portfolio of services

already available in modern EMRs. It is worth of

note the fact that while satisfaction for a feature was

an item that had been answered only by those that

had previously claimed to use an EMR already

ElectronicRecordsforGeneralPractice-WhereweAre,WhereweshouldHeadtoImproveThem

539

endowed with such a feature, the usefulness of a

feature was probed on the entire sample of

respondents, irrespective of the presence of such

feature in the respondent’s EMR.

Table 1: QUA ranking.

Rank Functionality Priority

1 Importing First ***

2 Backup First ***

3 Reminder First ***

4 Mobility First ***

5 SharingCA First ***

6 SharingGPs First ***

7 SharingColleagues First ***

8 Audit First ***

9 PubRetrieval First ***

10 Annotation First ***

11 SharingPatients First ***

As an aid to decision making and the process of

detecting risks and opportunities of either

developing from scratch or evolving specific

features, we also performed a ranking of the features

on the basis of the ordinal values collected in our

survey. Table 1 shows the result of this task that we

performed applying an original ranking method

purposely developed to infer what features are

valued most by the respondent sample. In this

method priority levels were assigned to single

features in the following manner: 1) we counted the

number of times each feature was ranked either first,

second or third (“first priority class”), or was ranked

in any other position ("second priority class"); and 2)

we assigned each feature to the priority level with

the highest number of occurrences. This assignment

was then tested with a Chi-squared test to check if

rankings could be due to chance and hence fail to

show a real difference in the frequency of

assignment to either the first or second class of

priority. As shown in Table 1, all features are to be

considered in the first priority class with a very high

confidence that this assignment is not due to chance.

However, the evaluations collected in our study

allows us to provide a ranking where data importing

capabilities, backup services and reminders are the

functions valued most highly, while the capabilities

of annotating any content in the EMR and to share it

with patients are those valued less.

3.4 Impact on Medical Work

Impact of ICT on medical work has been

investigated both in terms of quality of support of

specific steps of the medical process (QSM), and as

perceived change in known attributes of general

practice. Support quality was evaluated on a 5-value

ordinal scale, ranging from 1 (“very low quality of

support”) to 5 (“very high quality of support”).

Hence, for QSM the descriptive and inference

analyses are depicted in Figure 4. From this Figure,

it is clear that Diagnosis and Therapy are placed at

the opposite sides of the satisfaction spectrum: GPs

found “treatment” be managed properly by their

EMR, probably tapping in the vast amount of

research around order entry and drug prescription;

while they find support of diagnosis significantly

disappointing. Also the part of EMRs that covers

medical examinations and sign/symptom evaluation

is found to be an area that would benefit from efforts

towards improvement, e.g., like those envisioned in

(Cabitza et al., 2013).

In regard to the impact of EMR on GPs’ work,

this is clearly perceived as greatly more

documented, more informed, more documented,

more interesting and, most notably, much more

appropriate. This is the opinion of approximately 9

GPs out of 10, except for the dimension of

“interest”: in this case, 1 out of 6 GPs believes that

digitization has not changed the extent her work is

interesting; consequently, a lower number of GPs

deemed that their work has become more interesting

since the advent and diffusion of EMRs (76%).

Moreover, the majority of respondents indicated that

their work has become more stressful; however, this

finding is not associated with statistical significance

(43% more stressful, 41% less stressful; 17%

unvaried, Chi-squared=.62, df=1, P-value=.43, NS).

In regard to privacy concerns, the majority of the

respondents perceived their work as confidential as

it was before digitization (more confidential 39%,

vs. equal 43%, Chi-squared= .95, df=1, P-value=.33,

NS). This is an interesting finding especially in light

of the importance that usually IT researchers and

developers attach to the requirements of security,

privacy and confidentiality in the management of

health data. It would be interesting to investigate if

this perception of GPs is more related to a motivated

confidence in their tools, or rather to an actual

ignorance of the potential risks related to the

management and transmission of digital data.

3.5 Use of ICT to Communicate

In regard to the role of technology in mediating

doctor-patient communication, phone resulted the

communication means most frequently used: 95% of

the sample said to rely on phone to communicate

with patients at least “sometimes and with some

regularity” (79% said to do it often). Electronic mail

HEALTHINF2014-InternationalConferenceonHealthInformatics

540

Figure 4: Boxplots for the QSM items.

is the second tool in terms of frequency of use (52%

use it with some regularity, 19% often). Then SMS

(32% regularly, 9% often), fax (24% regularly, 4.9%

often); and finally (not surprisingly) Social Media.

This term was denoted in the questionnaire as any of

the following tools: Forum, Blogs, Chat and Social

Networks (like Facebook and LinkedIn);

interestingly, it resulted that these media are used

regularly by almost one tenth of the sample (9.8%,

2.8% often). In particular, 83% of the respondents

claimed to use emails to communicate with patients,

while almost 1 GP out of 5 (19%) claimed to use any

of the social media defined above. The frequency of

use of emails, social media and SMS resulted to be

correlated with statistical significance (Email with

SMS, large correlation: alpha=.53, P-value<.001;

Email with Social Media, moderate correlation:

a=.36, P-value<.001; Social media with SMS

moderate correlation: a=.39, P-value<.001). The use

of Social media and emails to interact with patients

was slightly correlated with claimed IT skills: the

higher the latter ones, the more frequently these

tools were used (for both items, Alpha=.18, P-

value<.001).

When we asked the respondents about how they

would like to communicate with patients in the

future (obviously besides face-to-face encounters),

the majority of the sample declared to be fine with

current use of phone, fax, SMS and social media

(more precisely, phone: 63%, P-value<.001; fax:

54%, P=0.016; SMS: 54%, P=.048; Social Media:

45%, P-values are for Chi squared tests, with 1

degree of freedom, two-tailed). However, almost one

third (31%) of the respondents wished they could

use phone less frequently, and similar results hold

also for fax (39%), SMS (28%) and, notably, also in

regard to a relatively novel means like social media

(38%), which has apparently disappointed some

expectations, at least as facilitator of doctor-patient

interaction. Interestingly, in regard to electronic mail

the results are opposite: only one respondent out of

10 expressed the wish to use emails less in the

future; indeed, the majority of respondents said they

would like to use emails more often in the future

(49%); the difference with respect to those who are

fine with current use (40%) is statistically significant

(Chi squared=7.2, df=1, P-value=.007). This finding

would suggest that doctors would prefer to move

from a distributed voice-based interaction with

patients to a message-based interaction, that is that

they dread interruptions and volatile indications

more than the work overload implied by having to

respond to more emails from their patients, probably

also in virtue of the famous proverb “spoken words

fly away, written words remain”.

In 2011 the same question was asked to a

representative random sample of Italian GPs in a

survey reported in (Cabitza, 2012). Comparing the

present survey and the previous one, we observe that

in 2013 a greater proportion of doctors claimed to

exchange emails with their patients to discuss

health-related problems “often and regularly” (11%

vs. 19%); this difference is statistically significant

(Chi-squared = 7.9, df = 1, P-value = 0.003). In this

two-year time span, also the use of Social Media

proved to have improved (14% vs. 19%, Chi-

squared= 2.7, df = 1, P-value = 0.0495). Also with

respect to prospective use, we detected statistically

significant increases in the proportions of GPs who

claim to want to use emails and social media more

often: in regard to emails we observed an increase

from the 30% of 2011 to the 49% of 2013 (Chi-

squared =27, df = 1, P-value < .001); and in regard

to social media from the 12% of 2011 to the 17% of

2013 (Chi-squared=3.1, df = 1, P-value = 0.038).

4 CONCLUSIONS

In this paper we have presented the findings of a

country-wide survey about the current and

prospective use by GPs of the ICTs (mostly EMRs)

that support their daily practice. The vast number of

GPs involved, the items considered and the analysis

of the responses collected in this study make this

research valuable for a number of aims. First to get a

picture of the level of adoption and appropriation of

EMRs in General Practice: the latter dimension was

ElectronicRecordsforGeneralPractice-WhereweAre,WhereweshouldHeadtoImproveThem

541

addressed by two specific items, PerceivedFit and

PerceivedExploitation, and their influence on

satisfaction was proved to be relevant. Second, to

begin considering analytically how multiple quality

dimensions and appropriation are intertwined, and

discover, for instance, how investments in IT

training and IT skill improvement for GPs could be

reflected in higher exploitation rates, better fit of

EMRs with the GPs’ needs and higher satisfaction.

Last but not least, to detect areas of improvement

(see Figure 4) and to assign priorities and rankings

to features in order to both increase overall user

satisfaction (see the NP score mentioned in Section

2.1) and decide where to focus on to make EMRs

better tools (see Table 1). Notably, our survey also

addresses how relatively new media could impact

practice: to this regard, attempts to include current

social media in medical practice by innovators and

early adopters (10% of the target population) seem

related to growing scepticism and disillusion.

However, GPs seem to still value written interaction

with their patients (cf. the increasing trends for email

actual usage and intention to use), and this could

hint at more communication-oriented models for the

next EMRs to come, as argued in (Cabitza and

Gesso, 2014)

In a period of fast and continuous innovation and

yet urgent spending limits to welfare and primary

healthcare, detecting the most value-adding features

of a class of applications for their reference key

users, and enabling the subsequent prioritization of

interventions to focus on could be a necessary move

to make ehealth a convincing driver for the feasible

progress of the medical profession.

REFERENCES

Benson, T., 2002. Why general practitioners use

computers and hospital doctors do not - Part 1:

Incentives. British Medical Journal 9, 1086-1089.

Cabitza, F., 2012. On the attitudes of GPs toward novel

features of their next EPRs. In: Quality of Life through

Quality of Information, Studies in Health Technology

and Informatics. pp. 911–916. Springer.

Cabitza, F., Simone, C., Colombo, G., 2013. ``Worth a

thousand fields’’. Arguing for a visual turn in

computer-supported general practice. In eHealth 2013:

Proceedings of the IADIS International Conference on

eHealth, 24-26 July, 2013 Prague, CZ, Computer

Science and Information Systems. IADIS Press, pp.

95-102.

Cabitza, F. and Gesso, I., 2014. Trying to fill the gap

between Persons and Health Records: the MedIcona

InterPersonal Health Record. In Healthinf 2014,

Proceedings of the BIOSTEC International conference

on Health Informatics.

Delaney, B., 2010. General practice at the cutting edge of

information technology, or failing to keep pace?

British Journal of General Practice 60, 239–240.

Dobrev, A., Haesner, M., Haesing, T., Korte, W.B.,

Meyer, I., 2008. Benchmarking ICT use among

General Practitioners in Europe. Gesellschaft fuer

Kommunikations- und Technologieforschung mbH.

Hassey, A., 2001. A survey of validity and utility of

electronic patient records in a general practice. British

Medical Journal 322, 1401–1405.

Hippisley-Cox, J., 2003. The electronic patient record in

primary care--regression or progression? A cross

sectional study. British Medical Journal 326, 1439–

1443.

Keiningham, T. L., Cooil, B., Andreassen, T.W., Aksoy,

L., 2007. A Longitudinal Examination of Net

Promoter and Firm Revenue Growth. Journal of

Marketing 71, pp. 39–51.

Misericordia, P., 2011. Indagine FIMMG sui modelli

organizzativi prevalenti negli ambulatori dei

generalisti d’Italia. Sanità - Sole 24 Ore 14, 14–15.

Porcheret, M., Hughes, R., Evans, D., Jordan, K.,

Whitehurst, T., Ogden, H., Croft, P., 2004. Data

Quality of General Practice Electronic Health Records:

The Impact of a Program of Assessments, Feedback,

and Training. Journal of the American Medical

Informatics Association 11, 78–86.

Purves, I. N., 1996. The paperless general practice. British

Medical Journal 312, 1112–1113.

Simon, S. R., Soran, C. S., Kaushal, R., Jenter, C. A.,

Volk, L. A., Burdick, E., Cleary, P. D., Orav, E. J.,

Poon, E. G., Bates, D. W., 2009. Physicians’ use of

key functions in electronic health records from 2005 to

2007: a statewide survey. Journal of the American

Medical Informatics Association 16, 465–470.

Swinglehurst, D., Greenhalgh, T., Roberts, C., 2012.

Computer templates in chronic disease management:

ethnographic case study in general practice. British

Medical Journal Open 2, e001754–e001754.

Treweek, S., 2003. The potential of electronic medical

record systems to support quality improvement work

and research in Norwegian general practice. BMC

Health Services Research 3.

HEALTHINF2014-InternationalConferenceonHealthInformatics

542