PRE2015 3 Groep3

Group Members

Student ID	Name
0903288	J.J.P. Beckers
0909421	N.J.A. Frints
0911264	R.G. Hup
0896239	S.J.W. Maas
0924760	H.V.C. Ramchurn
0923126	G.M. van Vliet

Pharmacy Service Help Robot

Nowadays, getting your medicine at the pharmacy takes more time than you have. Last years the waiting times at the pharmacy improved a lot. However, most people have a busy agenda these days. This results in the fact that the waiting times should be reduced even more. In addition, the society expects a 24-7 service. The current system cannot come up to these expectations.

To create a solution for this problem the idea of the Pharmacy Service Help Robot came up. The use of this robot will reduce waiting times a lot, because getting your medication via this robot will be a lot faster than via regular employees, especially in busy pharmacy. The robot will also be able to dispense your medicine any time of the day. The users will be able to retrieve the medicines that they need from the robot. This will apply both for prescription medicines and OTC (over the counter) medicines. Since most medicines require an explanation about the use, the robot will be capable to explain this to the users.

Not only will the system help the customers of the pharmacy, but it will also help the pharmacist themselfs. Taking over some parts of the current job of the pharmacist, results in them having more time available. With this they will be able to assign their time better and work in a more flexible way. It will also help with the increasing need for staff in the medical industry.

An important remark to this concept is that this robot will only be a tool that takes over a part of the pharmacist’s job. The Pharmacy Service Help Robot will NOT change anything to the current protocols.

Goal

Decrease waiting time

It is generally known that waiting times at pharmacies can be long. To verify this statement, literature research and both qualitative and quantitative research has been done.

According to NIVEL (Netherlands Institute or Health Services Research) , the average waiting time in a pharmacy, both public and private, is 9.2 minutes, with a standard deviation of 6.0 minutes. This includes queueing with other customers and waiting for the pharmacist to fetch the customers medicine. While a large part of the respondents (41,6%) states that they have to wait for 5 minutes or less, a significant part of respondents (44,6%) states that they have to wait for 10 minutes or more.

While these numbers represent both public and private pharmacies, the public pharmacies do have longer waiting times, with an average of 18 minutes. Respondents that are using public pharmacies stated that a waiting time of 7.5 minutes is most desirable. It could therefore be concluded that a significant part of pharmacy customers, both private and public, do not meet this desired waiting time.

As the average waiting time at public pharmacies it significantly higher than the average of both public and private pharmacies, it could be concluded that private pharmacies do have significantly shorter waiting times than public pharmacies. This could be confirmed by the dutch Mediq pharmacies, which state that their average waiting time is 1.77 minutes .

[ Results of our own research / questionnaire not completed yet ]

The conclusion could be made that the waiting time at pharmacies is acceptable in a lot of cases, but does not meet the desired values in other cases. While private pharmacies generally provide fast services, public pharmacies do lack such rapidity. By implementing the concept of a robotic pharmacy service help, the waiting times could be reduced, as a robot is faster than humans regarding fetching medicine, performing the majority of checks and dispensing the medicine to customers.

Research

Interview

We had decided to conduct an interview with a pharmacist, so we could expand our understanding of the current pharmacy system. We managed to get an appointment with Ellen Jansen, who works at Pharmacy Fellenoord. On the 24th of February two group members (Chiel van Vliet & Sil Maas) went over to the pharmacy and asked Ms. Jansen some questions. Both the questions and the (translated) answers can be found here: Media:Pharmacistinterview.pdf‎.

One of the most important findings of the interview was that Ms. Jansen stated that 2 minutes is about the minimum time required for a pharmacist to get the requested medicine to the customer. Most medicine take notably more time to deliver to the customer. Ms. Jansen explained that most of this time is spent on checks; scanning medicine package, checking contents of medicine package, etc. This means that if we manage to automate these checks we could greatly improve on the time it takes to give a customer his medicine.

We also found that the current system of explaining to a customer how his/her medicine should be taken and when it should be taken, is far from perfect. Ms. Jansen explained that there are a lot of foreign people who come to pharmacy Fellenoord and often they speak neither Dutch nor English. This means that when the pharmacist gives the obligatory written and verbal explanation of the medicine, he or she can only really hope that the customer understood this. This is another problem that our idea could solve, for instance with an interface that has different language options. This way, non-Dutch speakers can still understand the explanation, because it is given in their native language.

Mr. Jansen also mentioned that she had noticed that were a lot of budget cuts coming through in the pharmacy sector. Budget cuts can easily lead to a lack of employees, which would worsen the current conditions (longer waiting times). Our new pharmacy robot could possibly solve this as well, simply by being more cost-efficient in the long run. Eventually the pharmacies will have lower costs without jeopardizing the quality of their work.

After the interview we finally got a chance to look at the "behind-the-scenes" part of the current pharmacy "vending machine", which is only used for repeated prescriptions. We already knew that this system sends a code via an SMS message to the customer, which he/she can type in to the touch screen of the machine to get their medicine. The user-interaction part is outside of the building, like a cash dispenser. However, now we got to see how the medicine is put into the machine. The same checks have to be performed before the medicine can be put in the machine. After the checks are done, the pharmacists scans the medicine package with a scanner connected to the machine and the SMS message is automatically send to the customer. Then the pharmacy selects a size of the box she wants to store the medicine in and the machine grabs a box of the right size and holds it behind a small window. The pharmacy then opens that window and puts the medicine in the box and lets the machine put the box back in its place in the storage area. Ms. Jansen explained that at larger pharmacies they have too many repeated prescriptions, so they keep the medicine in the machine for a limited amount of time.

Questionnaire

Questionnaire results - pie chart of the waiting times

To gain more knowledge on the customer experience in pharmacies, we decided to create a questionnaire on the subject. The questionnaire had two multiple choice questions, one “yes/no” question and six open questions. By asking people to fill it in via social media we managed to get a total of 90 responses.

Our main finding during the analysis of the data was that 56.6% of the participants stated that they usually have to wait longer than 5 minutes before they got their medicine. Of that group, 14.4% usually wait longer than 10 minutes, sometimes even up to 30 minutes or more. This means that for the majority of pharmacies there is room for improvement concerning the waiting times.

Another interesting finding is that 83.3% of the participants state that they usually have a prescription when going to a pharmacy. This suggests that prescription medicine is the main reason for people to visit a pharmacy, meaning that changing the way pharmacies handle the sale of prescription medicine will have the biggest effect on the overall customer experience.

After giving a description of our Pharmacy Robot concept we asked the participants whether or not they thought it would help speed up the process of getting medicine at a pharmacy and whether they would use the system. 67% of the participants did in fact think our concept would improve the current pharmacy system, especially concerning the waiting times. Interestingly enough, only about 42% percent of the participants stated that they would want to use the Pharmacy Robot over the normal pharmacist (or assistant). Most of the people that did not want to use the Robot explained that they would not trust the Robot as much as they would trust a pharmacist (or an assistant) or that they would miss the explanation of the intended use of the medicine. The latter is simply caused by the fact that we did not give a very detailed explanation of our concept, because we were afraid it would bore people too much and that it would result in less serious answers and thus worse data. However, the lack in trust might be an actual problem, because 25.6% of the participants indicated that they would be afraid of the system malfunctioning and giving them the wrong medicine. However, this is a phenomenon that we have seen before, namely with the introduction OV-chip card system. There used to be tons of articles (like this article and this one) going around about the lack of trust in the system, whereas nowadays it is, for the most part, widely accepted as the current public transport system (as described in this article about increase of trust). So, maybe it is just a matter of giving people time to get used to the system and let their trust grow.

Trust

In order for our idea to be successful, we need the customers to trust the system. So, we did some research on subject. We found a lot articles relating to 'trusting automated systems', but only a few that actually contained conclusions leading to specific concepts that we might be able to incorporate into our system.

Our main finding is that the level of trust people have in system is mainly determined by the performance of that system on a task (Hancock et al., 2011) (van den Brule et al., 2014). This means that if the pharmacy robot is consistently faster than a pharmacist (or assistant), while still being able to give the customer the necessary information, the customers will trust it easier. Also the environmental situation appears to have some influence on the level of trust. This might be a good reason to have the user interaction part of the system located inside the pharmacy (like we intended in the first place), as opposed to the current repeated prescription system, of which the UI is outside of the pharmacy. There you, as a customer, are standing on the street, with people walking by you constantly. Inside the pharmacy it will be much quieter and much more pleasant, and thus increase the level of trust in the system.

A commonly used strategy to increase the public trust in a new system is to let the concept be known before implementation (Hengstler et al., 2016). By letting people get used to the idea of this new system and with the use of good PR, the trust can grow before the system is actually used. Also, by inviting some people to try out the system, you can not only get useful feedback, but also have people spread the word about their personal experiences with the system. If done right, this could have a major positive influence on the trust in the system.

Another very interesting finding is that the more anthropomorphic features (name, gender, voice, etc.) the system has, to more people expect it to do well on a task (Waytz et al., 2014). In this particular research participants drove an autonomous car using a driving simulator. One car was just able to control the speed and the steering, whereas the other one also had some anthropomorphic features. "Behavioural, physiological, and self-report measures revealed that participants trusted that the vehicle would perform more competently as it acquired more anthropomorphic features." So, the more an automated system (mentally) resembles a human, the more trust it gains from its users.

USE

Users

The primary user of the pharmacy robot will be the pharmacy's customer and the employees of the pharmacy, since they both get an operable screen. The secondary users will be the doctors, since they won't have to worry abouy their patients forgetting important information (because this is stored on their personal cards). The tertiary users will be the manufacturers of the machine, because they will build, deliver and probably install the machine as well.

For the users the robot will be a great improvement of the current pharmaceutical system. They will have to wait less and they will be able to get their medicine and the explanation for it in private and in their native language.

Society

The effects that the pharmacy robot will have on society as a whole are not very clear, mostly because a lot of people rarely visit pharmacies to begin with. One could argue that if people have to wait less in pharmacies, they will get happier and therefore increase the overall societal happiness, but this is a little far-fetched.

Enterprise

For the pharmaceutical companies it will probably be a very good investment, because it will lead to lower costs.

Implementation

Pros and Cons Table

The most important points while taking the implementation of this concept into account are things such as identity verification, issuance of non-prescribed medicines and the human assistance required to supervise and ensure the good functioning of such a system. These points are crucial for the optimum advantages that this system is supposed to bring forward. Working through the concept as a project though do bring certain pros and cons into consideration:

The system

In the current state of the art in the pharmaceutical industry there are only a few automated systems. The rest is still done (for the major part) by humans. One of the automated machine currently in use, is a machine to pick up your repeated prescription. It basically is a big vending machine outside of the pharmacy where you are able to pick up your medicine. Once your repeated prescription is at the pharmacy, the pharmacist performs all the regular checks and then puts the medicine into the machine. Next you will receive a SMS code to retrieve your medicine from the machine.

Since there are a lot of improvements to be made with automated systems, our project will focus on doing that. Now the storage of the medicine is all done manually. Everything is put into place by the pharmacists and once they need a certain medicine they will go to the storage and retrieve it themselves. Because we want to expand the current system of the repeated prescriptions to a bigger system, the whole storage of the pharmacy will be automated. All the medicines will be stored in an automated storage system. Multiple interfaces will be present for either the pharmacist to use or for the customer getting their medicine. The pharmacist will be able to store the medicines in the systems once they are delivered to the pharmacy and be able to retrieve all of them freely again. The user will be able to retrieve their repeated prescriptions as well as general medications like aspirin and bandage via another machine present at the pharmacy.

Flowchart of the system

We have seen that not all checks performed by the pharmacist are able to be automated. When the pharmacist has to open the box of the medicine to check if it is the correct medicine and if it is the correct dose, is one of these checks which (with currently present technologies) is not possible to perform by robots. Therefore we have chosen to automate as much as is possible with current technologies. This has resulted in the separation between the two different uses of the new system.

The flowchart (on the left) shows how the general process within the pharmacy will work with our newly developed system.

As can be seen, the storage is shared for both uses. Once the medicines are delivered, they are quickly checked (by opening the boxes) and then put into storage. Now a couple of different things can happen.

If a customer is going to the pharmacy to retrieve a medicine for the first time or a medicine where tight checks have to be performed, they will have to go to the present pharmacist. But now instead of having to retrieve the medicine requested themselves, the pharmacist can simply ask the machine to retrieve the specified medicine for them. While the machine is doing this, the pharmacist is able to ask the customers all questions necessary and check whether or not he is allowed to use it. Once this has all been done, the machine will have delivered the medicine and the pharmacist will only have to check if it is the correct medicine and has the right dose. This will reduce the time for these situations.

For the other situation; if a user calls for their repeated prescription, the pharmacist will use the same system to retrieve the medicine and will perform the regular checks they do now. Once this has all been done, the medicine will be put back into storage but with an extra (digital) label, stating that it is a repeated prescription for that particular person. Once this has all been done, the user will get a message saying that he or she is able to pick up their medicine. To make sure that the system can validate the identity of the user, a membership card of the pharmacy will be used. Once the user comes to the pharmacy to retrieve their repeated prescription, they will simply walk up to the machine and check their membership card. Once this is checked, a screen will popup saying that the user has a repeated prescription present, ready for pickup. Now the user can select to retrieve the medicine. He or she will also be given the option to buy extra medicines. A menu will appear where the user is able to select the medicines they require. If a certain medicine is selected, a screen with info about the general use of the medicine will appear and the user will have to confirm that this is indeed the medicine they wanted. If confirmed, a number of question will be asked regarding the current state of the user (if required for giving the user that specific medicine). Once these questions have been answered, the medicine will be added. Now the user will have to pay for all their medication and once this is done, the machine will retrieve them all from storage. Finally the locker/dispenser at the bottom of the machine will open and the user can get all their medicines.

From the above it can be concluded that the complete system of the Pharmacy Service Help Robot will consist of multiple different parts. The major parts are: the storage system, the user interface and the controllers. The system will physically be split up in the different parts. We will especially focus on the implementation of the user interface(s). The other parts are already developed or / and can be developed in the future.

Storage

There already exist automated systems capable of retrieving packages from a storage area. Since these technologies already exist, we will not focus on this in our project.

<Maybe we could mention some of these systems over here>

User Interface

User Interface

The user interface is the part of the system that we will be focusing on. It needs to abide by certain criteria:

• User friendly to all age groups
• Understandable and clear information
• Touch screen or easy and well defined buttons and command points
• Available in most useful languages
• Able to interpret same medicine compositions but with different business or marketing names
• Able to scan customer/membership cards

The global cliënt-interface will be something like this:

• Step 1: Scan your card
• Step 2: Interface (screen) interaction
  • Screen 1: Language selection
  • Screen 2: Medicine selection
  • Screen 3: General information about the medicine
  • Screen 4: Return to screen 2 or continue (when all the medicines are selected)
  • Screen 5: Questions
  • Screen 6: Payment
• Step 3: Get the medicines from the dispenser

The different drop screens according to the different options chosen by the users have comprehensible and straight forward to make the people more confident in using it; especially when dealing with health related issues, people might be wary of inserting or pick the wrong option or information. The color theme, as we consider it, is more advisable to be kept green and white (As shown in the picture), as these are the colors that people usually associate with pharmacy already nowadays.

Also a pharmacist-interface will be needed, because the pharmacist will also need to interact with the system. However, for this project we will focus on the cliënt-interface of the system.

Controllers

From the flowchart it can be seen that multiple controllers will be needed to make the system work. Multiple checks are always performed to make sure that the user gets the right medicine. One of the most important things is to check whether the user is not allergic to the medicine or is already taking other medicine which can cause problems if taken together. In our system this will be done by checking the database of the user and asking the user questions via the user-interface. In the database the physical state of the user and all the medicines the user is already taking, need to be listed for the system to work. However, the questions are also important, because a database cannot be trusted 100%. By comparing all this information, the system will be able to check if the user is allowed to receive the medicine.

Another important controller is the one checking if the medicine gathered from the storage is the correct medicine. This controller will have to check if the (contents of the) box correspond to the requested medicine. The pharmacist will check before the medicine is placed into the storage if the box and the medicine inside correspond. When this is correct, the pharmacist will add a label to the medicine, specifying which medicine (+ the dose of the medicine) is in the box. This makes it a lot easier for the system to check the medicine, since it is hard for a machine to open a box and look into it. If the medicine has been retrieved from storage, the controller will check via the barcode on the label if it is the same as requested by the system. If this is the case, the medicine can be given to the user and if not, another attempt will be made to retrieve it.

Smaller checks to see whether the user has selected the correct medicine and if the user understood the explanation of how to take the medicine are simple and will not need an advanced controller to work.

Results

To design the user interface, we have chosen to use a website as a proof of concept. Our work in progress can be followed on the following web page: Mecine Dispenser

Conclusion

conclusion of the project

Planning

The Gantt-chart for a good overview of the planning

On the right a Gantt-Chart is shown that we made in the beginning of the project to get a good overview of the planning. We tried to stay with it as much as possible.

Reflecting the planning

In week 1 the concept is devised. In the first instance, the problem had to be defined. When this was known, we came up with the solution for the problem: the concept. After developing the concept a little, research was required into i.a. current systems. The problem, the concept and the research needed to be presented in presentation 1 (week 2).

After presentation 1 we received feedback about our concept etc. The feedback had to be processed in week 2, before presentation 2 (week 3). An important part of the feedback was that the research should go deeper to support the problem better. Unfortunately, we could not finish this before presentation 2 because we wanted to interview a pharmacist and we wanted to distribute questionnaires. These things, however, took more time than expected. Also the concept had to be devised more and needed to become clearer.

During processing the feedback we already started to implement the robot. We also got the instructions for the wikipage after presentation 1, which resulted in the fact that we needed to start working on this as well. We started doing this dedicated in week 3. The feedback of presentation 2 did not get any better. As result, we needed to do a lot in week 3 to convince the teachers that the problem was a real problem and that the concept was great.

In week 4 we had to describe the concept detailed, because during the meeting on Monday it became clear that this was not done enough. This was the most important thing that we needed to do this week. Furthermore, we started working a little on the implementation of the concept. We started with investigating how we could implement our concept. We had done this to start implementing in week 5.

Files

ToDo -> add file links

• Concept presentation
• Feedback presentation