Analysis of a “Simple” Medical Device
Contents
1. Figure 1 A blood glucose meter and associated components WINTER 2001 ERGONOMICS IN DESIGN 7 Surun uros s19919p V U9UA dr ns 3s 3 Jo uontl1u rro Jodoid Apsessoour dis 3s 13 33 sur dwers Ayddy s amp e dsrp 1333W Jopjoy drs 4533 ot Jo uontoo Jopjoy dins 3s 1 ot QUT dims 3s94 M SU LTT jxe 003 drns sysasut ejds p uo SUONINAISUT 4J sqo JOU SI0q suononssut s e dsrp 1339W Avydsip Jo uonzoo drn s Wosuy LS 03 IIU IOJ LM 9 Z Z qyoyoed u do o qeun IPL qyoyoed Jo uonevoT aseyoed oy wos dis 3593 e aAOTIOY T Z surpvar sey sAvydstp Jajour woqNq 33820 JOUUe uo p uim u uA sd q 199W uoynq JJO uO Jo uonto2o T IPU I uo WINE ZZ p xrur moq ur Ajrodoid 30u uonnjos JoRUo tu yeus JOU JETA XIUI uonnjos IRL ainposoid 499 1107 AJSNOIOSIA RIA uonnjos OUOD IP LYS Z Z ela uonn os p sn uonnjos jo zpuo2 posdxy uoN aep uonestdxa Jo uonzoo Ionuojo ya uo ap UONRAIdxs ay YoY TTT p sn Sur q st uonnjos p sn uonnjos O QUOD 7931109U uoN ad4q uonnyos 1991109 o nuoo sooni 8 3991102 IY WYL PAJO TTT 3S9 uonnjos o nuo2 sooni 8 z WIO A Z 9S9 IOJJLJSNES Jo uontorpur IOJJLJSNES SI I AfLIOA put INSI dAJSGO O1 reg ynsor uonzsauqt e3 sAe dsrp 133W Avydsip ay Jo uonLvoT J9J9U u1 uo Surpe 9A49SqO OL T Z peystuy usym sd q ainpasoid 4991109 UMOP JUNOD 0 1979W 10 VEM JOU S20 q pue unos sdeydsip 1939W ejds p 3y Jo uonsooT 0 0p Woy UMOP JUNOD 0 19W P dOJ eM2. The system also provides very lit tle feedback about the accuracy of performance and the consequences of errors are high Other devices in this class would include blood pressure monitors heart rate monitors oxygen tanks and infusion pumps In this article we demonstrate how human factors ergonomics tools can be applied to medical devices in general and blood glucose meters in particular We focus on blood glucose meters because of their widespread use and availability but our approach and findings can be gen eralized to other home health care devices 6 ERGONOMICS IN DESIGN WINTER 2001 REGAN H CAMPBELL amp ARTHUR D FISK Definitely Not As Easy as 1 2 3 Tt s as easy as 1 2 3 begins the videotape that accompanies a commonly used blood glucose meter Simply set up the meter check the system and test your blood However our analysis of the system sug gested that it is not quite that easy Sure there are three general steps involved in using the system but imple mentation of those steps requires the user to do 52 sub steps The first column of the table on pages 8 9 lists all the steps required to test one s blood using a typical blood glucose meter Setting up the meter requires 6 steps checking the system requires 22 steps and test ing of the blood glucose level requires 24 steps This level of complexity is not unique to one particular meter we conducted a task analysis on another com mo
3. GTZ dis yooyo e ueya 19430 Surly WOS 79979p 1 p ou dr ns 3s93 oy Zup wos JSUT Apoess1o0our drns JOI SUJ U UIM UMOP SUNOD 1919 4 1 p ou dins 3s 1 Jo uoneooT ozur drys JOP u1Jo Z PIS PJS ZTT ainpasoid 1991109 SUONIN ASU IYI 10J HEM JOU SIO suononssut s e dsrp 1339W Avydsip un Jo uont 2o aC OPIS Masur Avs 0 199W oI IOJ VEM LTZ pasow st dis u UA Jopjoy digs yoyo Iaou Jou s20 q T DIS H SUJ s ejdsp 1339W ainpasoid 20 digs 3s 1 t Jo mo dims yoayo oy IPS Q T Z suononnsur JO JEM JOU SIOP tuonnjos O QUOD JO aanpas01d 7991109 poolq s rjdde sapjoy woy dims yay saou Jou s oq suononmsur skedstp 1939W Avydstp oy Jo uonzoo a dures Ayddy Avs o 132w ay OJ IPA STZ digs 3 u e ucu JayI0 SUPIWOS s1231 p W u UA dims yooys Jo uonequario r odoid Jopyjoy dins 3s 3 oy olur Surun utos 311 sur Apn ro our drns yoyo 311 su a dures Ajddy sAe dsrp 139W fI pjoy dr ns 3s 1 Y Jo uoncoo drns yoyo 9u3Jo J pis piIS PTZ jxe 003 dins sir sur Avydstp uo suon nzsur sArasqo 10U s oq suononssur s 4e dsrp 1339W Ae dsip Jo uonzoo ANG Wsu Avs 03 1339W OJ PEM ETZ Supeo asep s amp e dsrp raqour uoynq 37220 JouUR_D uo poun u ym sd q 199W uonnq JO uO Jo uontoo uo WU p um ZZ p uce Jou LI ISIL uoN LIIL 3s 1 Jo UONLIOT u 9 SI LITE 3897 OYA AINS JEW TTZ 3s 1 dims JOIO e wod Tz WS S oy yoy oz Avydsip uo sasueys apo JOQUINU 9pOd PAIOUN 19IUY UON uonnq 7 IRL uonnq O IP JO uon
4. entry and analysis of the surveys and Richard Pak and Aideen Stronge for their constructive comments This work was supported in part by a grant from the National Institutes of Health National Institute on Aging Grant No P50 AG11715 under the auspices of the Center for Applied Cognitive Research on Aging one of the Edward R Roybal Centers for Research on Applied Gerontology W
5. p Surouvg Apo zro our poseydor Jo poorydas jou deo orA p Surpurg posou jou deo aanoa101g A J991JOIUT psu JIUL T poaouras jou deo 321p Sururg suonon sut JO VEM JOU SD0 Apsessoour dS 3S9 JSUT j e 003 drns sosur ejdsrp uo suon on nsur A1 sqo 1ou sooq yoyoed uado o qeun uoynq 397290 JOUTL 4ep uonesdxa YPIY O IVT p uc o9 JOU LIW ISIL WY JIIIOOUT JOYIVS fU UL 1398104 IFUL 7991109 I UI SI 3I JII A pue INSI IAI SqO 07 eJ JNS 4 3S9 1ou 0 Ie UMOP UNOD 0 1979W U1 107 HEM JOU SIO uonvoo suoIM o pordde uonnjos pasn uonnyjos 1593 1o5u SUOTIIN AISUT I0 WM JOU S90q uoN uoN 3ue q soos Avydstq a ns 3s 1 s e dsrp 1339W p usrug u uA sdaoq pue unos sXe dsrp 1939W p 152 1 p MSTOU u uA aou0 sdaaq 199W p npo d poojg papd 1328Uy J297 IPL R UON SuoN p220 UIM SJON Mp SUNULT ln ur ln ur ln ur ln ur suononnsur s ce dsrp 1939W SUIPOUIOS 8799P H u UA dwers Ayddy s amp e dsip 1333W suono nzsui s e dsrp 1339W uoN surpeas sey sAvydstp Jajour uo poun uym sd q 1939W UON UON uoN UON a ns 3s 1 s ejdsIp 1939W poysruy u uA sd q pue yuno s ejdsIp 1939W injsrout S1J991 p 1 u UA sd q 1939W suononasut skedsip 1939W ainpsso01d 7393110 3 ns 3s93 JOQuIDUTOY uopnq JO uO I Jo uontoo sy ns 3S9 Jo uontorpur Avydsip t Jo uontzoo ainpas
6. portion of the meter i e the display or the test strip holder As a result aspects of the procedure may be missed Usability Analysis To assess whether viewing the video instructions would provide sufficient training for someone to per form the tasks we tested a sample of individuals on sev eral of the component tasks We focused on the three calibration tasks set the code perform a check strip test and perform a glucose control solution test Six participants four older adults mean age 70 years and two younger adults mean age 21 years watched the portion of the video relevant to the three calibration tasks and then performed each calibration test one at a time It is not appropriate to blame the user for making an error when the root cause of the error may really be the design of the system itself lations of text complexity based on a transcription of the video instructions yielded a rating of about the seventh grade level comprehensible by 72 of the population This indicates that the vocabulary may be more ad vanced than desired to ensure comprehension by the general population Analysis of the video is instructive It begins with an introduction and advertisement recommending the product This is followed by a testimonial from a phar macist who recommends the meter and then a brief overview of the major steps required to perform a blood glucose test Following the overview each of the main steps proc
7. substeps A variety of problems might occur when testing blood glucose levels In some instances the problems may not affect the test result For example if the user forgets to verify whether the code number on the meter matches the number on the test strip vial there are no consequences But sometimes errors may affect the blood glucose reading For instance the code on the meter may differ from that on the test strip vial The user may neglect to perform regular check strip or con trol solution calibration tests Or the user may not wait for the prompts on the meter and may be unable to obtain a test result at all The most critical errors are likely to be those that result in an incorrect glucose reading which could then lead to the incorrect insulin dosage or erroneous adjustments in other treatments Perhaps the most striking aspect of the task analysis is the complexity of this supposedly simple medical device With 52 total steps for calibrating testing and using the meter there are many opportunities for error The sequential nature of the task itself might contribute to problems because an error early in the process car ries through e g inserting the strip incorrectly Given the complexity of the task it is crucial that users receive the necessary instructions to be able to use the system safely and effectively Instruction Analysis Another mechanism for understanding the potential sources of errors is to review
8. the type s of instruction that the user receives Perhaps the instructions are too complex for the user to understand or the instructions 10 ERGONOMICS IN DESIGN WINTER 2001 A ee ees WS gt Y aS S2 2 Figure 2 Top Obtaining a blood sample bottom applying sample to a test strip The strip is then inserted into the meter We analyzed the instructions that were provided with the meter for reading level using the Flesch Kincaid Grade Level analysis The user s manual was written at an eighth grade level and would be readable by about 58 of the U S population The instructions for using the lancing device were written at a sixth grade level and considered readable by 72 of the population However the instructions for the test strips were written at a more advanced level ninth to tenth grade which is read able by only 51 of the population In the United States about 23 million people aged 25 and older would not be able to read and understand materials written at this grade level U S Census Bureau 1997 Therefore we determined through the reading level analysis that the instructions would not be understood by segments of the intended user population In our analyses we also considered the content of the instructional video provided with the meter Calcu video also switches back and forth between showing the actor in the segment and a close up view of the meter or the relevant
9. 0 id 490 1109 Avydstp oy Jo uonzoo jdures poo q Ajdde 01 uonesoy sodoig ainpasoid 19941079 uoynq ISLI Jo uontzoo Josuy Jord 0 uontoo 1994107 urn Jo yysusy Jodoid ainpasoid 4991107 ainpasoid 4991107 3914P SuIUL IP X202 o1 MOH des so1aap SuDpuc verdor o3 MOH des An32391o4d AOuI91 0 MOH JOUR 31 sur 037 MOH ainpasoid 1994107 ainpasoid 4991109 Avydsip oy Jo uonzoo dis 3s 3 Jo uonezuano Jodoid roppoy drns 3s33 ot Jo uontzo2o T Avydsip Jo uonzoo yoyoed Jo uonevoT uoynq JJO uO Jo uontoo T Hep uonestdxa Jo uontoo LITE 3s 1 JO UONRIO T sway posmbar zo uoneso posmboas ore suai PYM aanpaso id 791109 adeyord uo IBULI 3231102 I JO UONRIO T s NsaJ 3s 3 Jo uontorpur t e dsip oy Jo uont32o ainpasoid 7991109 Avydstp oy Jo uonzoo uonnjos Jonuos dde 0 19UA t Inp 2o d 199110O ainpasoid 4991109 Aejds p oy Jo uonzoo preosip pue so1aop Suroury IY WOT JODUL IYI IAOW qnsai 3593 INO pio5 q He PPVU oY UN syNsot 4S9 asoonys poo q tn pey 091 Sb WOI UMOP JUNOD o1 IOPU IAP 1OJ VEM dims 3saj 03 dutes pooyq Ajddy pooyq jo dosp sursuey adrey e ure3qo 0 JaSuTy oY z nb uonndq SL ay SSAI Josuy B JO opIs oY surede 9014p SUIOU IY POE spuoo s 01 JO opts Ino Je ue INOA BULH spury nod useAA rtA 9p Suroury IYI 3509 deo sotaap Bunuej ot 2t d g des sanoaioid 3 2uc IYI JJO 3SIA 1 p ou 395ut IYI
10. 39 646 Moray N 1994 Error reduction as a systems problem In M S Bog ner Ed Human error in medicine pp 67 91 Mahwah NJ Erlbaum Tomky D M amp Clarke D H 1990 A comparison of user accura cy techniques and learning time of various systems for blood glu cose monitoring The Diabetes Educator 16 483 486 U S Census Bureau 1997 Educational attainment of people 25 years old and older by sex March 1997 On line Available http www bls census gov cps pub 1997 educ_att htm Van Cott H 1994 Human errors Their causes and reduction In M S Bogner Ed Human error in medicine pp 53 65 Mahwah NJ Erlbaum Ward W K Haas L B amp Beard J C 1985 A randomized controlled comparison of instruction by a diabetes educator versus self instruction in self monitoring of blood glucose Diabetes Care 8 284 286 Wickens C D Gordon S E amp Liu Y 1998 An introduction to human factors engineering New York Longman Wendy A Rogers is an assistant professor Regan H Campbell is a doctoral student and Arthur D Fisk is a professor in the School of Psychology Georgia Institute of Technology Atlanta GA 30332 0170 wr43 prism gatech edu Amy L Mykityshyn is employed by the Georgia Tech Research Institute We thank Marsha McNatt of the Diabetes Association of Atlanta for information and guidance in this project Anar Desai and Heather Harrison for their assis tance with data
11. Analysis of a dimple Medical ayl 2 EDICAL DEVICES ARE DESIGNED TO improve the health of an individual Advances in the technology of medical devices may enable patients to take an active role in maintaining their own health One example is the hand held battery operated meter that patients with diabetes can use to check their blood glucose levels Through self monitoring diabetes patients can deter mine their blood sugar levels and adjust diet insulin or exercise to effectively manage their diabetes The advent of self care unfortunately has some negative consequences as well The patient must take primary responsibility for monitoring blood glucose levels without assistance from a health care profession al As a result there may be an increase in the occur rence of errors in the use of blood glucose meters see Laux 1994 for monitoring difficulties when meters are unavailable These errors can have serious conse quences for the short and long term health of the indi vidual particularly because treatment is often determined from glucose level readings Diabetes can lead to many complications if not properly treated including high blood pressure heart or kidney disease vision problems or blindness complications in pregnancy and death Blood glucose meters represent a class of home health care technologies in which the tasks are sequen tial in nature that is each step is reliant on the success of the previous step
12. a noninvasive unit Recent design improvements on blood glucose meters could address some of the aforementioned broad cate gories For example the category modify the blood sampling procedure may have been addressed with recent blood glucose sampling advances However other design issues remain and the new systems may themselves have unforeseen problems that require human factors ergonomics analysis One design improvement evident from our task analysis see the table on pages 8 9 is the need for appropriate task specific feedback For example the meter should be able to recognize a check strip used for calibration versus a test strip used for glucose testing After recognizing the test strip the meter should pro vide the feedback that is appropriate to the task the user is performing In addition given the sequential nature 13 ERGONOMICS IN DESIGN WINTER 2001 meter that is small enough to carry around but when the system gets too small it becomes difficult to work with and the display cannot be easily read among other things The critical point is that design ideas must be tested with the target user population Instructional Design Nearly 50 of the respondents in our survey said the user s manual was a primary source of instructions Given this reliance on training materials it is critically important that such materials be well designed Ideally the development of training programs and instructional materia
13. ding is accu rate for the particular set of test strips The second task was to perform a check strip test Four people two younger two older successfully com pleted this test One older adult was unable to perform the task initially but when prompted to start over she was able to complete it successfully The remaining older adult performed the test by happenstance when trying to do something else The third task was to perform a glucose control solu tion test Four participants two younger two older were able to obtain a control solution reading One older adult received an error on her first try and was not able to complete it a second time The remaining older adult was not able to get a reading because he failed to use a test strip on his first try and even when prompted he could not complete the test It is important to note that no one remembered to check the type of control solution or to check the expiration date on the solution even though these activities were described on the videotape Participants had some difficulties performing the three calibration tests that represent only a small subset of the tasks required to use a blood glucose meter suc cessfully see the table on pages 8 9 If they had watched the entire video and then been asked to perform all the tasks we suspect they would have had even more diffi culties Note also that the older adults tended to have more difficulties but performance was not perf
14. ect for the young adults either In sum the video instructions may appear to be an organized and straightforward means of training someone to use a blood glucose meter but our empirical evidence suggests otherwise The task analysis presented in the table illustrates the complexity involved in using a blood glucose meter Clearly it is not a simple three step process There is a lot to learn and the multitude of sequential steps leads to the possibility of a cascade of errors caused by both design and training issues Proper design of the device coupled with clear and easy to follow instruc tions is critical In addition the design of the system and of the instructional materials must meet the needs of all potential users meter is not trivially easy to use and requires some experience to use it comfortably More than 70 reported having problems using the system when first learning to use it They reported problems such as remembering the steps setting up the meter calibrating the meter using the lancet getting a blood sample and reading the display Perhaps more telling was the finding that survey respondents had used an average of 2 5 different brands of meters range 1 7 mode 3 Dissatisfaction with particular blood glucose meters led them to try new meters although there were also cost reasons for switch ing Reasons for changing meters were as follows 40 wanted a meter that was easier to use 25 sought differ en
15. edures for calibration testing blood glucose levels and cleaning maintaining the meter is described by a different actor The pharmacist concludes the video by instructing users to refer to the user s manual for further instructions The manufacturer s toll free customer support number is then provided On the surface the video may seem useful for train ing it provides demonstrations of each component and every critical step is described and shown with a close up view of the meter and other equipment as someone performs the step In some instances editing tech niques such as a grayed out screen with the item of interest highlighted are used to focus attention on the pertinent element in the scene However our instruc tional analysis revealed some potential problems First the main steps are never reiterated or reviewed in the video Only when the pharmacist provided an overview of the basic steps to test blood glucose levels were the instructions previewed or summarized The I ERGONOMICS IN DESIGN WINTER 2001 Only two of the participants one younger one older were able to set the code on the meter The remaining four skipped the first calibration altogether and tried to use a check strip test One of these partic ipants a younger adult realized that he had not set the code and later performed this calibration The other three all older never checked or changed the code which is important in ensuring that the rea
16. entially encountered and the anecdotal comments are of in terest Respondents were asked how many times they had to use their blood glucose meter before they felt com pletely comfortable with it 32 said 1 2 times 40 said 3 4 times 20 said 5 10 times and 8 said they had to use the meter more than 10 times in order to feel comfortable This pattern shows that the blood glucose 12 ERGONOMICS IN DESIGN WINTER 2001 least in part from the user s manual Most users learned to use these systems on their own and likely believe that they are performing the task adequately However self instruction based on manufacturer manuals may be insuf ficient for accurate performance with a blood glucose meter Ward Haas and Beard 1985 found that after 30 minutes of self instruction and practice participants made errors 20 to 40 of the time The next most common way for respondents to learn to use a blood glucose meter was from a nurse doctor or health professional 30 However these individuals may not have the time to devote to training users for some meters participants in another study needed an average of 23 9 44 5 minutes to learn to use the system and they continued to make errors even after this training Tomky amp Clarke 1990 The remaining respondents in our survey learned to use their meter from other materials provided by the manufacturer 8 trial and error or past experience 8 or a friend or family memb
17. er 8 Note that several people provided more than one answer to this question and the percentages represent the percentage of the total number of responses Toward Solutions for Reducing Error Based on prior studies our task analysis reported problems from current users and observed difficulties calibrating a blood glucose meter users clearly have dif ficulties using these devices and there are many sources of error If they are unable to use meters properly patients are unable to adjust treatment of their diabetes There are two areas where improvements can reduce errors the design of the system and the design of instruc tional materials that teach people to use these devices System Design Our survey data provide some directions for system redesign We asked respondents If you could design a new and improved blood glucose meter what things would you change Their answers covered these broad categories 1 Modify the strips make them longer standardize the test strips reduce the price make the area on the strip that receives blood larger automatically indicate an in sufficient amount of blood and allow an addition of blood to the same strip change the packaging of strips of the tasks involved meters that provide prompts for the proper sequence of the tasks might help to reduce errors see Gardner Bonneau amp Gosbee 1997 for additional design recommendations It is important to consider the benefit
18. es errors resulted from a general lack of care on the part of the patient in complying with the manufactur er s instructions p 803 However we would be wise to heed the counsel of experts in the field of human error To assign blame to an individual who makes an er ror is no assurance that the same error will not be made again by a different person Van Cott 1994 p 61 People of good intention skilled and experi enced may nonetheless be forced to commit errors by the way in which the design of their environment calls forth their behavior Moray 1994 p 67 In other words it is not appropriate to blame the user for making an error when the root cause of the error may really be the design of the system itself or the type of instructions provided to users Uncovering Sources of Errors How do we determine the actual sources of errors in a particular system The field of human factors ergonomics has much to offer in this regard The strength of an HE E approach is that it can be predictive in helping to anticipate errors as well as prescriptive in motivating design and instructional development Proper HF E analysis can reveal when the fault might lie with the design of the system or instruction rather than with the user Moreover human factors analysis can reveal ways to improve design and instructions to minimize error Task Analysis The path to understanding the source of errors in a task begins with an in depth
19. is misleading When testing a drop of blood using a normal test strip the user sees the message Apply Sample on the meter and this feed back is useful However if the user is calibrating the meter using a check strip and Apply Sample appears he or she is supposed to slide the check strip out of the test strip holder In this circumstance the feedback message is misleading and potentially error inducing were poorly designed or the users never received any instructions to start with Home medical devices often have step by step instruc tional manuals The meter we examined came with three instructional pamphlets a user s manual instructions in cluded with the test strips and instructions for a lancing device which holds the lancet i e sterile needle used to prick the finger to obtain a drop of blood for testing The meter also included an instructional video We analyzed the content of the text materials and the instructional video to determine whether these materials might con tribute to user error A global measure of readability of text material is the grade level and percentage of the population that could read and understand it Standard writing equates to a seventh or eighth grade level but the recommended level of writing for general information is sixth grade McLaughlin 1969 Sure there are three general steps involved in using the system but implementation of those steps requires the user to do 52
20. ls should follow Instructional System Design models for a review see Wickens Gordon amp Liu 1998 Such models include an analysis to determine the specif ic training needs for the system A task analysis such as we conducted would be included as well as an analysis of the characteristics of the trainees e g education perceptual and cognitive abilities and of the resources available e g access to a videocassette recorder for video based training Design and development of the training materials involves selecting the method of training e g step by step instructions modeling and the media in which the training will be administered e g text video For both written and spoken instructions particular care should be taken in the following areas Ensure that readability levels are appropriate for the intended user population Keep the vocabulary simple and explicit avoid jargon Minimize the need for readers and listeners to draw inferences Use numbered steps to describe procedures Emphasize critical components or warnings Provide advance organizers such as an initial over view or outline of the information to be presented to help users conceptualize the information Describe the steps in the same order in which they will be performed Provide redundancy of information either through multiple modes text and figures written and spoken or through the repetition of critical informati
21. more serious for classes of devices that were initial ly designed for hospital use but are now being used in the home such as ventilators and infusion pumps Klatzky amp Ayoub 1995 We caution development teams to be aware of the potential for design induced errors to make every effort to design out those errors and then to devel op adequate training and instruction protocols for the users of the device 14 ERGONOMICS IN DESIGN WINTER 2001 References Colagiuri R Colagiuri S Jones S amp Moses R G 1990 The quali ty of self monitoring of blood glucose Diabetic Medicine 7 800 804 Gardner Bonneau D amp Gosbee J 1997 Health care and rehabili tation In A D Fisk amp W A Rogers Eds Handbook of human fac tors and the older adult pp 231 255 San Diego CA Academic Kelly R T Callan J R amp Meadows S K 1991 Impact of new medical technology on user performance In Proceedings of the Human Factors Society 35th Annual Meeting pp 699 702 Santa Monica CA Human Factors and Ergonomics Society Klatzky R L amp Ayoub M M 1995 Health care In R S Nicker son Ed Emerging needs and opportunities for human factors research pp 131 157 Washington DC National Academy Press Laux L 1994 Visual interpretation of blood glucose test strips The Diabetes Educator 20 41 44 McLaughlin G 1969 SMOG grading A new readability formula Journal of Reading 12 8 6
22. nly used meter and found 61 substeps When people are told that a system is trivially easy to use but have difficulties with it they tend to blame themselves They may be too embarrassed to ask for assistance stop using the system altogether or contin ue to use it without realizing they re using it incorrect ly Additional evidence that blood glucose meters are not trivially easy to use comes from a study conducted to assess the accuracy of the blood glucose readings obtained by diabetics using blood glucose meters Colagiuri Colagiuri Jones amp Moses 1990 Colagiuri et al tracked 90 patients using two different meters for one month Participants were asked to test previously measured solutions The results showed that 62 of the patients made at least one error that was classified as clinically significant meaning that the patient either would have taken a medically inappropriate action or failed to take a medically appropriate action based on the incorrect reading According to the error classification conducted by Colagiuri et al the most frequent cause of the error was a faulty technique on the part of the user In these researchers words the most commonly encoun tered error arose from the patient not adhering to the Contrary to manufacturers claims about ease of use real users report a number of problems with complex devices at home specific instructions of the manufacturer I Jn most cas
23. olur JIDULJ IJS L 11 su deo ra p Sur ut ay DAWA pooyg JO doip BIOL a duies Ayddy Avs 07 1339W at 10J Ue Jopjoy dins 3s 3 at QUT dims 3s 3 HSU INS ISU LS 03 IIU 10J PEM aseyoed oy Woy drs 3533 IAOUY IPU IY UO UMJ aseyord drns 3533 un uo 1ep UONRIIdxe ay JYO ULI st varie dys 3s 1 t ANS aq o2 yoy Quer o rA p Suroury sdins 480 19 9U S RLIOIVUT 3991103 IP 1919 poolq mod 3s I 8ue3o ed drns 3s oy uo 8uti IYI UA IPU IP uo Surpeas IP edur sqnsot 189 uonnyos JOonuoo omy peed 0 9 p wo uMop UNO 0 I 919uln IP JOJ HEM drys 3s3 03 uonnjos Jo NUOD asoonys Ajddy a duies Ayddy Avs o3 1332W at 10J HEM STE bre ere cle Ive Ore ore 66E SOE LOE OE SOE t 6 COE COE TOE OE se Le 9 E S E ve CE TE T E OE CVO Ive Orez OCT STC 9 IN DESIGN ERGONOMICS WINTER 2001 buttons to change the time on the meter recall read ings stored in memory and perform calibration proce dures In all meters reviewed the tasks were sequential and required a great deal of procedural knowledge to obtain accurate blood glucose levels Whether the system provided information for each step of the process tended to vary In most cases feed back from the meter was not informative The meter provides the same prompts regardless of the task being done a calibration procedure or actual glucose testing Sometimes the feedback
24. on Training materials should be developed iteratively in other words materials are tested with typical users of the systems benefits of training are assessed and the mate rials are refined as needed In the case of training and in struction for a blood glucose meter testing must include diabetics who represent the range of education percep tual function and cognitive ability levels of the typical diabetic who will be expected to use the system The last phase of instructional system design is pro gram evaluation Is the training system effective Do the training materials enable trainees to use the system safe ly and effectively The evaluation should be based on users across time tracking errors they make determin ing the potential need for refresher training e g do they lapse into bad habits forgetting critical steps such as cal ibrating the meter and so on Conclusion The possibility for design induced human error asso ciated with blood glucose meters is prevalent which argues strongly for human factors intervention Our anal ysis focused on a single device but a blood glucose meter is representative of a class of medical devices Our purpose was to demonstrate the need for human factors ergonom ics intervention early in the design process of medical devices especially those designed for home use The blood glucose meters we reviewed were designed specifically for patient use The HF E concerns may be even
25. s of ostensible design improvements For example early meters required that excess blood be wiped off the reagent strip Newer no wipe technology eliminated this aspect of the procedure However eliminating this step did not improve the blood sample because errors still occurred if the test strip was not sufficiently covered with the sample Kelly Callan amp Meadows 1991 So although the improved technology may have minimized certain kinds of errors it did not solve the problem of obtaining an adequate blood sample It is also important to consider the constraints of design recommendations Design involves trade offs For example users may want a portable blood glucose Most users learned to use these systems on their own and likely believe that they are performing the task adequately 2 Modify the meter reduce the amount of program ming required make it simple and easy to use make steps more accessible make it smaller 3 Modify the features add memory date and time shorten processing time increase memory include attachable clamp so it can be worn like a beeper 4 Modify the blood sampling procedure require a small er amount of blood develop better ways to get a drop of blood 5 Modify major systems eliminate the need for calibra tion make it self contained design a blood glucose meter that can communicate directly with an insulin pump make it lighter build a lancet device within the system develop
26. t features 15 were seeking a lower cost alternative and 20 had received recommendations from others When asked how they had first learned to use a blood glucose meter 49 of respondents said they learned at Participants had some difficulties performing the three calibration tests that represent only a small subset of the tasks required to use a blood glucose meter successfully Participants Self Reports Another valuable tool in the HF E arsenal is to interview the users of a particular system We conduct ed a survey of people who have experience using blood glucose meters The survey included questions con cerning demographic information e g age education sex race blood glucose meter usage patterns e g fre quency assistance length of experience difficulties using the system e g remembering steps reading the display and open ended questions about likes dislikes and suggestions for improvements Surveys were completed by 26 individuals with a mean age of 59 years 11 women 15 men 4 African American 21 Caucasian 1 mixed Anglo Hispanic The average number of years of formal education was 15 or more years and all respondents reported having at least some college education All were diabetics who used a blood glucose meter typically at least once per day and had been using a meter for an average of seven years Although this sample is relatively small the data are suggestive of the range of problems pot
27. too T you sapoo IP pun uonnq O IP ssd E TI aseyord dims 3s 3 pue 1979W aseyord uo Jaquinu po2 3991109 puy JouURD_ uoN JOqUINU po2 1991109 Jo UONLIO T ap UO sioqumu apod sy aIeduIOD Zz ZI surpeas sey sAvydstp Jajour uoynq 3320 JOUUe uo poun u uA sd q 199W uoynq JO uO jo uontoo IPU oy uo UMJI ZI 1932W dq poO TI uoynq 33220 JouUR_ uonse uonnq J29 IPL wong O Jo uontoo uonng O IP sLpPA E TI uonynq asvajat o1 reg Surpeod sey s e dsrp Joqour uopnq 23e90 JOUR uo pausing uaya sd q 199W uonnq JO uO Jo uonto2o uoynq gO uO P SL pue ssatg TTI uoynq 3320 3ouurO uona WONG J297 IPL uoynq O Jo uoNRO T uoqng O y pjoy pue ssaig TTT asensury Aefdsip IP PAPS TI IW oy dn 395 OT swia qo1g enu loq ypeqpso7 syu ur asmb g ystL ON aS popMOUYL yse 3seL ULLAW ASOOATH adoo ltd GuvaNvLlg V AOA SISAXTIVNY ASV LE ERGONOMICS IN DESIGN WINTER 2001 uonkoo ofesun UT JOOURT PABOSIp JOUR dAOUTOI 0 9 qtuQq YNsat p1OIII O rez 3 ns 1593 SUOIM p4Ao5o q JJO IPU UNA 0 IE T 3J nS 3S9 9JOU 07 e4 UMOP JUNO 0 1979W JYI JOJ VEM JOU SIO Surpeas e 10 p r dde poo q ysnous Jou uoneooy Jadoid uo posed jou poo g Suisury ucu Joyies poresurs sr poo q paonpoid poo q ysnoua oN Josuy yond 01 qeuq poolq 398 02 qgun 28u ysurege p 9u JOU orA9p unue si 8uu 03 MOT poojq poos 308 0 qeun opis w Wwe suey 1 Uu0q uono Jur qtssod paysem jou spueH p 3 o2o Jou ra
28. understanding of what a person has to do when performing that task see Figure 1 page 10 We conducted a task analysis of a sample blood glucose meter to help us understand the errors that may be made when using a typical meter see the table on the following pages The meter we analyzed is readily available at major pharmacies and is representa tive of commonly used systems It requires three cali bration procedures First each time a new box of test strips is used a code on the meter must be set to match a code provided on the test strip vial Second the meter must be calibrated periodically by using a special check strip Third to ensure that the meter is providing cor rect glucose readings the patient is instructed to per form regular tests using a glucose control solution Each task was defined in terms of the information required by a user to complete each task task knowl edge requirements the feedback provided by the sys tem and the potential problems that might arise if the task were not carried out properly see the table Test ing blood glucose levels using a meter requires substan tial procedural step by step knowledge The task analysis demonstrated that many of the tasks require knowledge of the correct procedure and location and function of the control button to successfully complete each step This particular meter has two control but tons whereas other meters we examined had three B gt yaqa I
Download Pdf Manuals
Related Search