University of Toronto scientists use machine learning to fast-track drug formulation development

New study demonstrates the potential for machine learning to accelerate the development of innovative drug delivery technologies

Peer-Reviewed Publication

UNIVERSITY OF TORONTO - LESLIE DAN FACULTY OF PHARMACY

 

IMAGE: [LEFT TO RIGHT] CHRISTINE ALLEN AND ALÁN ASPURU-GUZIK FROM THE UNIVERSITY OF TORONTO ARE COMBINING EXPERTISE IN PHARMACEUTICAL SCIENCES, AI AND MACHINE LEARNING TO DEVELOP NEW DRUG FORMULATIONS FASTER. view more 

CREDIT: STEVE SOUTHON

TORONTO, January 10, 2023 [Embargo 10 a.m. London Time] – Scientists at the University of Toronto have successfully tested the use of machine learning models to guide the design of long-acting injectable drug formulations. The potential for machine learning algorithms to accelerate drug formulation could reduce the time and cost associated with drug development, making promising new medicines available faster.

The study was published today in Nature Communications and is one of the first to apply machine learning techniques to the design of polymeric long-acting injectable drug formulations.

The multidisciplinary research is led by Christine Allen from the University of Toronto’s department of pharmaceutical sciences and Alán Aspuru-Guzik, from the departments of chemistry and computer science. Both researchers are also members of the Acceleration Consortium, a global initiative that uses artificial intelligence and automation to accelerate the discovery of materials and molecules needed for a sustainable future.

“This study takes a critical step towards data-driven drug formulation development with an emphasis on long-acting injectables,” said Christine Allen, professor in pharmaceutical sciences at the Leslie Dan Faculty of Pharmacy, University of Toronto. “We’ve seen how machine learning has enabled incredible leap-step advances in the discovery of new molecules that have the potential to become medicines. We are now working to apply the same techniques to help us design better drug formulations and, ultimately, better medicines.”

Considered one of the most promising therapeutic strategies for the treatment of chronic diseases, long-acting injectables (LAI) are a class of advanced drug delivery systems that are designed to release their cargo over extended periods of time to achieve a prolonged therapeutic effect. This approach can help patients better adhere to their medication regimen, reduce side effects, and increase efficacy when injected close to the site of action in the body. However, achieving the optimal amount of drug release over the desired period of time requires the development and characterization of a wide array of formulation candidates through extensive and time-consuming experiments. This trial-and-error approach has created a significant bottleneck in LAI development compared to more conventional types of drug formulation.

“AI is transforming the way we do science. It helps accelerate discovery and optimization. This is a perfect example of a ‘Before AI’ and an ‘After AI’ moment and shows how drug delivery can be impacted by this multidisciplinary research,” said Alán Aspuru-Guzik, professor in chemistry and computer science, University of Toronto who also holds the CIFAR Artificial Intelligence Research Chair at the Vector Institute in Toronto.

To investigate whether machine learning tools could accurately predict the rate of drug release, the research team trained and evaluated a series of eleven different models, including multiple linear regression (MLR), random forest (RF), light gradient boosting machine (lightGBM), and neural networks (NN). The data set used to train the selected panel of machine learning models was constructed from previously published studies by the authors and other research groups.

“Once we had the data set, we split it into two subsets: one used for training the models and one for testing. We then asked the models to predict the results of the test set and directly compared with previous experimental data. We found that the tree-based models, and specifically lightGBM, delivered the most accurate predictions,” said Pauric Bannigan, research associate with the Allen research group at the Leslie Dan Faculty of Pharmacy, University of Toronto.

As a next step, the team worked to apply these predictions and illustrate how machine learning models might be used to inform the design of new LAIs, the team used advanced analytical techniques to extract design criteria from the lightGBM model. This allowed the design of a new LAI formulation for a drug currently used to treat ovarian cancer. “Once you have a trained model, you can then work to interpret what the machine has learned and use that to develop design criteria for new systems,” said Bannigan. Once prepared, the drug release rate was tested and further validated the predictions made by the lightGBM model. “Sure enough, the formulation had the slow-release rate that we were looking for. This was significant because in the past it might have taken us several iterations to get to a release profile that looked like this, with machine learning we got there in one,” he said. 

The results of the current study are encouraging and signal the potential for machine learning to reduce reliance on trial-and-error testing slowing the pace of development for long-acting injectables. However, the study’s authors identify that the lack of available open-source data sets in pharmaceutical sciences represents a significant challenge to future progress. “When we began this project, we were surprised by the lack of data reported across numerous studies using polymeric microparticles,” said Allen. “This meant the studies and the work that went into them couldn’t be leveraged to develop the machine learning models we need to propel advances in this space,” said Allen. “There is a real need to create robust databases in pharmaceutical sciences that are open access and available for all so that we can work together to advance the field,” she said.

To promote the move toward the accessible databases needed to support the integration of machine learning into pharmaceutical sciences more broadly, Allen and the research team have made their datasets and  code and available on the open-source platform Zenodo.

“For this study our goal was to lower the barrier of entry to applying machine learning in pharmaceutical sciences,” said Bannigan. “We’ve made our data sets fully available so others can hopefully build on this work. We want this to be the start of something and not the end of the story for machine learning in drug formulation.”

 

 

   - 30 -     

 

Media Contact:
Kate Richards
Leslie Dan Faculty of Pharmacy, University of Toronto
kate.richards@utoronto.ca
(416) 978 - 7117
 

About the Leslie Dan Faculty of Pharmacy, University of Toronto
The Leslie Dan Faculty of Pharmacy at the University of Toronto is Canada's top-ranked faculty of pharmacy, offering cutting-edge undergraduate and graduate programs. We are globally recognized for impactful pharmaceutical sciences research and fostering expert and innovative clinical practice. Our scientific research focuses on the role of pharmacists in the health care system, and the full scope of drug discovery and delivery. We advance education programs that develop leaders in science and clinical practice and work to strengthen the link between research, education, and patient care.

 

About the Acceleration Consortium, University of Toronto
Based at the University of Toronto, the Acceleration Consortium (AC) is a global community of academia, government, and industry who are accelerating the discovery of materials and molecules needed for a sustainable future, from renewable energy and biodegradable plastics to resistance evasive antibiotics. The AC builds self-driving labs that combine artificial intelligence (AI), robotics, and advanced computing to radically reduce the time and cost of bringing these materials to market.

---

JOURNAL

Nature Communications

DOI

10.1038/s41467-022-35343-w 

METHOD OF RESEARCH

Experimental study

SUBJECT OF RESEARCH

Not applicable

ARTICLE TITLE

Machine learning models to accelerate the design of polymeric long-acting injectables

ARTICLE PUBLICATION DATE

10-Jan-2023

Low concern and political distrust behind vaccine-resistance, new study finds

Peer-Reviewed Publication

UNIVERSITY OF KENT

With a return to the workplace and school, the UK Health Security Agency recently warned that cases of flu and COVID-19 are expected to soar throughout January. Currently, it is estimated that one in eight NHS beds are taken up by flu and COVID-19 patients – yet 22 million vaccines for these viruses have not been used.

A new paper, looking at the psychological reasons for people’s unwillingness to accept the COVID-19 vaccine, reveals two key factors to be people’s lack of concern for the consequences of contracting the virus together with their lack of faith in the government's actions.

Since the outbreak of the coronavirus pandemic in 2020, substantial efforts have been directed toward the development of effective vaccines.

The success of national vaccination campaigns is considered central to finally containing the virus and finding a way out of the pandemic. Yet, as seen this January with 9 million people who are still due to take a COVID-19 booster jab, and 13 million unused free flu shots, vaccine efficacy and safety are not enough to determine the success of these campaigns - vaccine acceptance among the public is also key. This is why it is so crucial to understand psychological reasons for vaccine hesitancy.

The study, conducted by Professor Dominic Abrams from the School of Psychology at the University of Kent, and Dr Fanny Lalot (University of Kent and University of Basel), tested the theory that COVID-19 vaccine hesitancy is a result of ‘distrustful complacency’ – a dangerous combination of low concern and low trust. The psychologists hypothesized that either concern or political trust should be sufficient to motivate people to get vaccinated, as the presence of one can compensate for the absence of the other. The absence of both concern and trust, however, ‘distrustful complacency’, would undermine that motivation, resulting in greater vaccine hesitancy.

Across two studies, 9,695 respondents from different parts of Britain reported their level of concern about COVID-19, their trust in the UK government, and their intention to accept or refuse the vaccine. Across the studies, respondents with both low trust and low concern were 26%–29% more hesitant to receive the vaccine, compared to those with both high trust and high concern.

The study showed that people who accept the vaccine do so because they are highly concerned about the consequences of the pandemic - for themselves and for others. Others do so because they trust the political institutions responsible for enacting the vaccination program. However, those who, for whatever reasons, do not trust these institutions and are also not concerned about the virus are much likelier to be hesitant about vaccination.

Professor Abrams said ‘Vaccines save lives but only if people are willing to take them. Our evidence shows that science and politics are powerfully interconnected. For scientific knowledge to make its contribution, people need to trust that politicians are using the evidence and advising wisely.’

Dr Lalot added ‘The importance of finding the same connection between trust, concern and vaccination intentions amongst White British, Muslim, and Black respondents, and generally across all nations in Britain. This underlines the importance of psychological perceptions regardless of demographic factors.’

ENDS

The paper ‘Distrustful Complacency and the COVID-19 Vaccine: How Concern and Political Trust Interact to Affect Vaccine Hesitancy’ is published on Wiley Online Library. The research is funded by Nuffield Foundation, and is part of a collaborative project ‘Beyond Us and Them’, between University of Kent and Belong – The Cohesion and Integration Network.

Dominic Abrams is a Professor of Social Psychology and the Director of the Centre for the Study of Group Processes in the School of Psychology at the University of Kent. He was recently awarded an OBE for services to social science in the New Year’s Honours list. His research examines all aspects of relations between different social groups and the behaviour of groups in general. Professor Abrams has extensive experience in the areas of equality and human rights, prejudice, discrimination, social attitudes and social change across the life course.

Dr Fanny Lalot is a postdoctoral researcher at the University of Basel and honorary scholar at the University of Kent. Her research interests revolve around social influence and behaviour change, motivation and goal systems, identity, group systems, and social and political trust.

About the University of Kent

The University of Kent is a leading UK university producing world-class research, rated internationally excellent and leading the way in many fields of study. Our 20,000 students are based at campuses and centres in Canterbury, Medway, Brussels and Paris.  

We are renowned for our inspirational teaching and our graduates are equipped for a successful future allowing them to compete effectively in the global job market.  

We are committed to supporting outstanding research and innovation across the arts and humanities, sciences and social sciences.  Our discoveries and research will emphasise existing and new signature areas, where we match the best in the world.   

The University is a truly international community with over 40% of our academics coming from outside the UK and our students representing over 150 nationalities.  

We are a major economic force in southeast England, supporting innovation and enterprise and through collaboration with partners, work to ensure our global ambitions have a positive impact on the region’s academic, cultural, social and economic landscape.  

---

JOURNAL

Political Psychology

DOI

10.1111/pops.12871 

METHOD OF RESEARCH

Survey

SUBJECT OF RESEARCH

People

ARTICLE TITLE

Distrustful Complacency and the COVID-19 Vaccine: How Concern and Political Trust Interact to Affect Vaccine Hesitancy

Metal-free batteries raise hope for more sustainable and economical grids

Peer-Reviewed Publication

KING ABDULLAH UNIVERSITY OF SCIENCE & TECHNOLOGY (KAUST)

 

IMAGE: ILLUSTRATION OF THE HIGH-EFFICIENCY METAL-FREE BATTERY DEVELOPED BY KAUST RESEARCHERS. UNLIKE CONVENTIONAL BATTERIES, THIS BATTERY COMBINES AN AMMONIUM-CATION-CONTAINING ELECTROLYTE WITH CARBON-BASED ELECTRODES. view more 

CREDIT: © 2022 KAUST; HENO HWANG

Rechargeable batteries that use ammonium cations as charge carriers could provide ecofriendly and sustainable substitutes to metal-ion-based batteries, researchers at KAUST show.

 

Metal-ion batteries, such as lithium-ion batteries, are the go-to energy storage solution. They dominate the market for portable consumer electronics and electric vehicles because of their high energy density and versatility. However, metal ions used in the electrolytes come from limited and declining resources, which threatens long-term availability. Their toxicity and flammability can be unsafe and harmful to the environment.

 

There have been several attempts to generate ammonium-ion-based batteries to solve sustainability and environmental issues because these cations are lightweight and easy to synthesize and recycle. However, ammonium cations are prone to reduction into hydrogen and ammonia at low operation potential, preventing the batteries from achieving their full potential. They also dissolve readily in electrolytes, making them difficult to incorporate into electrode materials.

 

Husam Alshareef, postdoc Zhiming Zhao and coworkers developed a high-efficiency metal-free battery by combining an ammonium-cation-containing electrolyte with carbon-based electrodes. The graphite cathode and the organic semiconductor anode are cheap, environmentally friendly and renewable, Zhao says.

 

With the ammonium cations, the researchers chose hexafluorophosphate ions as negative charge carriers and exploited the ability of graphite to reversibly accommodate these anions within its layers to create a “dual-ion” battery. In the battery, cations and anions simultaneously insert into their corresponding electrode during charge cycles and are released into the electrolyte during discharge cycles.

 

This differentiates our work from other studies, Zhao says.

 

“We designed an electrolyte that is both antioxidative and antireductive by screening a series of solvents resistant to high voltage and also taking into account its reduction stability,” Zhao says.

 

The antioxidative solvent mainly solvated anions participating in the cathode reaction, while its antireductive counterpart formed a solvation sphere around cations involved in the anode reaction. “This configuration is crucial for battery stability,” Zhao explains.

 

The battery outperformed existing ammonium-ion-based analogues with a record operation voltage of 2.75 volts. “It is now possible to develop high-energy nonmetallic ion batteries that can compete with metal-ion batteries,” Zhao says.

 

The team is currently working to enhance the performance to get closer to large-scale applications. “We are exploring anode materials with a higher capacity, which is crucial for improving the energy density,” Zhao says.

 

Alshareef’s group is developing cheap alternatives to lithium-ion batteries, particularly for grid-scale storage. “To eventually completely decarbonize the grid, the battery costs must significantly come down”, says Alshareef. Replacing lithium with nonmetallic charge carriers, such as ammonium ions, can help lower these costs. 

---

JOURNAL

Angewandte Chemie International Edition

DOI

10.1002/anie.202212941 

ARTICLE TITLE

A 2.75 V ammonium-based dual-ion battery

Emergency remote teaching during COVID-19 lockdown brought families closer together, but also required a lot from parents

Peer-Reviewed Publication

UNIVERSITY OF EASTERN FINLAND

School closures and emergency remote teaching caused by the COVID-19 pandemic put a particular strain on families with children with special educational needs. Yet, most parents felt competent in supporting the learning and remote schooling of their children, and they also enjoyed the increased freedom and autonomy, which made it possible to structure daily life according to their family’s needs and preferences. However, parents were nevertheless also concerned about the learning, well-being and relationships with friends of their children. 

Emergency remote teaching forced families to tackle a new kind of challenge, as parents had to balance between the demands of their work and those of their children’s remote schooling. A new study among 120 parents in Finland, conducted by the University of Eastern Finland and the University of Turku, focuses on parents’ perspectives regarding their competence, autonomy, and relatedness, in relation to the schooling of their children with special educational needs.

Parents hoped for better teacher-parent interaction

Parental well-being is crucial to positive parent-child interaction, as well as to supporting child well-being. This is why the researchers wanted to pay attention to the strengths and challenges parents experienced regarding emergency remote teaching of their children with special educational needs at the beginning of the lockdown in Finland. 

“Earlier studies have shown that emergency remote teaching and lockdowns had a major impact on the life of children with special educational needs, and on their families. For instance, the availability of therapy, rehabilitation and mental health support was often be reduced during lockdown. Children reacted to these changes with their behaviour, since there were also challenges related to children’s basic needs, such as social interaction, sleeping and eating,” Assistant Professor Kaisa Pihlainen of the University of Eastern Finland says.

Maintaining routines and essential relationships is crucial for many children with special educational needs.  In some children, breaking routines caused anxiety, depression, or rule-breaking or aggressive behaviour.

“When emergency remote teaching began, parents felt that they did not receive adequate support from the school or from teachers, and home-school interaction was inadequate especially for children with special educational needs. According to parents, teachers did not ask about their views on things, nor enquired about the workload caused by remote schooling.”

According to Pihlainen, it seems that maintaining and offering different forms of support affect not only the learning of children, but also the well-being of their parents.

Hectic school environment replaced by more peaceful learning at home

Although most parents reported that they were able to support the learning and remote schooling of their children, flexibility on the part of their employer was needed. Some parents felt that their work kept piling up or was done poorly as a result of providing support for their children: after all, some of children needed constant parental support for their learning.

“From the perspective of family dynamics, remote schooling brought children and parents closer together. Instead of the hectic daily life of schools, parents were able to create their own routines at home, and children, for example, enjoyed their meals better than at school. In families with two parents, shared responsibility also supported parental well-being in the new situation.

Remote schooling also enabled the educational content to be broken down into smaller bits, allowing children to progress at their own pace.

Towards increasingly individualised learning?

Pihlainen says that the results of the study confirm earlier findings. The pandemic did not treat families equally: parents with a higher level of education were better equipped to support their children, whose learning wasn’t hampered by remote schooling.

“In the future, finding ways to strengthen all parents’ capacity to support their children’s learning as part of the basic mission of the school is something to look into.”

According to Pihlainen, the benefits of remote teaching should also be reaped. The traditional school environment can be noisy and restless, in contrast to the possible peace and quiet of home. Being able to progress at one’s own pace and having the opportunity to concentrate support well-being and promote learning.

“We could consider how these observations can be made use of in the everyday life of schools, as well as on any future periods of remote teaching. Could our schools in the future offer calmer environments for learning and enable a more flexible and individualised pace of learning? Technology could also be better utilised, for example in the case of children’s travel or long-term illness,” Pihlainen concludes.

---

JOURNAL

European Journal of Special Needs Education

DOI

10.1080/08856257.2022.2159280 

ARTICLE TITLE

Parents’ competence, autonomy, and relatedness in supporting children with special educational needs in emergency remote teaching during COVID-19 lockdown.

Subway stations near river tunnels have worst air quality

Peer-Reviewed Publication

NYU LANGONE HEALTH / NYU GROSSMAN SCHOOL OF MEDICINE

Subway riders waiting in stations near tunnels that run below the city’s rivers are exposed to higher levels of hazardous pollutants than are found in other stations. The “river-tunnel effect,” as researchers call it, may help explain extremely poor air quality in the nation’s largest underground transit system and have particular implications for stations close to rivers in general.

In a previous investigation of New York’s subways, researchers at NYU Grossman School of Medicine found considerable variation in air quality among city subway stations. While some had pollutant levels a few times higher than that of outdoor air, the air quality in others was comparable to sooty air contaminated by forest fires or building demolitions.

To better understand why, the NYU Grossman research team measured air quality samples in 54 NYC stations during morning rush hour. They found that stations neighboring river tunnels had 80% to 130% higher concentrations of potentially dangerous particles in the air compared with stations only two or three stops further away from rivers. The new study published online Dec. 30 in the journal Transportation Research Part D: Transport and Environment.

“Our findings help explain why some underground subway stations are more polluted than others,” says study lead author David Luglio, MS; a doctoral student at NYU Grossman School of Medicine. “Those subway stations closest to rivers clearly must be prioritized during cleaning efforts.”

To explain the “river-tunnel effect,” Luglio notes that while many tunnels in the city’s underground subway system have some degree of air exchange with the surface, those traveling beneath water have more limited ventilation. As a result, harmful debris gets trapped and builds up over time. Trains passing through may then throw these iron and carbon particles back into the air and push them into the closest stations — those at either end of the tunnel.

The investigation, which Luglio says is the largest exploration to date of how river tunnels influence air quality in underground subway stations, also revealed that proximity to a river tunnel was the strongest factor in predicting a station’s pollution levels, followed by its age. Other potential contributors, such as station size and depth, did not appear to play a major role in air quality differences.

The Metropolitan Transit Authority reported that 5.5 million people rode New York City’s subways every day in 2019, before the COVID-19 pandemic began. According to past research, passengers were exposed to air with high levels of particles, which experts have linked to lung and heart disease as well as overall higher risk of death.

For the investigation, researchers collected over 100 air samples in stations between February and March 2022. Among the results, the study showed that on average, pollutant levels in all measured stations exceeded the daily exposure limit advised by the Environmental Protection Agency, which assesses potential health hazards in the environment.

For comparison, and to confirm the river-tunnel effect, the study team measured particle buildup on the B-line, a train route that crosses the East River via a bridge instead of passing beneath the water. Notably, pollutant levels in the two stations closest to the river on this train route were lower than that of stations farther away — as expected, the reverse of the river-tunnel phenomenon.

“Now that our results have identified key contributors to poor air quality in New York City’s underground subway stations, we have a better idea of where to improve conditions in the most contaminated areas of the transit system,” says study senior author Terry Gordon, PhD. “Increasing ventilation and scrubbing the tunnel walls and floors to remove continually recycling debris may make stations safer for riders and transit workers,” adds Gordon, a professor in the Department of Medicine at NYU Langone Health.

Gordon, also a member of NYU Langone’s Center for the Investigation of Environmental Hazards, cautions that since the investigation only explored subways in New York City, it remains unclear whether the river-tunnel effect occurs in other cities as well.

He adds that the study team next plans to examine the effects of subway contaminants on human cells to better pinpoint the level of exposure needed to pose a risk to human health.

Funding for the study was provided by National Institutes of Health grants ES000260 and ES007324. Further funding was provided by the NY/NJ Occupational Safety and Health Center ERC Pilot Project Award grant T42 OH008422.

In addition to Luglio and Gordon, other NYU Langone study investigators involved in the study were Tri Huynh, BS; and Antonio Saporito, BS.

---

JOURNAL

Transportation Research Part D Transport and Environment

DOI

10.1016/j.trd.2022.103579 

METHOD OF RESEARCH

Observational study

ARTICLE TITLE

Investigation of a river-tunnel effect on PM2.5 concentrations in New York City subway stations

Award stickers and taste descriptions matter for artisanal cheese buyers, Oregon State research shows

Peer-Reviewed Publication

OREGON STATE UNIVERSITY

 

IMAGE: EXAMPLE OF IMAGE SHOWN TO CONSUMERS FOR OREGON STATE UNIVERSITY STUDY. view more 

CREDIT: OREGON STATE UNIVERSITY

CORVALLIS, Ore. – Consumers are willing to pay more for familiar, versus unfamiliar, varieties of cheese if there is a sticker on the cheese indicating it won an award or if sensory information about the cheese – such as a description of its taste or food pairing suggestions – is included, a new study from Oregon State University shows.

The study also identified two broad groups of consumers whose cheese buying preferences differ. A group that prefers unfamiliar foods is willing to pay a premium for unfamiliar cheeses and an award sticker plays a much more important role than sensory information. The opposite is true for consumers who prefer familiar cheese varieties: sensory information play a much stronger role in willingness to pay more.

The study was in part motivated by the shift to online grocery shopping, which accelerated during the COVID-19 pandemic. While online grocery shopping has its notable advantages, the researchers note, the impacts can vary greatly for different food categories.

For example, specialty food products such as wine or cheese that are made on a small scale and have traditionally relied on in-person recommendations or product sampling might be more in need of new marketing strategies because online shopping can’t provide a real-time, in-person tasting experience.

“This is an under-studied area that is growing in importance, especially as people shift to buying groceries online and as subscription food boxes grow in popularity,” said Nadia Streletskaya, an assistant professor of applied economics at Oregon State. “Our study can help specialty food producers, many of whom operate on a small scale with limited budgets, determine best ways to promote their products.”

The researchers expect that the patterns they found with artisan cheese consumers could hold for buyers of other specialty foods, such as wine or different milk types, but more research is needed to make that conclusion.

For the study, the researchers evaluated how sensory information and the presence of award labels affected consumer demand for two familiar (brie, cheddar) and two unfamiliar (Coulommiers, Cantal) varieties of artisanal cheeses in the U.S. A total of 488 artisanal cheese consumers from two regions – 270 from Corvallis, Oregon and 218 from Ithaca, New York – took part in the online study.

Participants were shown side-by-side images of two cheese varieties, with price information as well as some combination of an award sticker or sensory information about the cheese. An example of the sensory information, this for the Cantal: “A tangy and bold cheese with a crumbly, hard texture.”

The award sticker and sensory information were chosen because they are common and relatively low-cost promotional strategies that translate well to the online retail environment.

After being shown the images, participants were asked to select which cheese they preferred to purchase. They also had an option to make no purchase.

The researchers found participants fell into two broad groups:

• The group that prefers unfamiliar foods, which made up about 44% of the total, look for cheeses not known to them and display a significantly higher willingness to pay for them. The researchers found that such consumers already are willing to pay a premium for less familiar varieties and an award sticker and sensory information further increase their willingness to pay.
• Consumers who don’t appreciate unfamiliar varieties, who accounted for about 47% of the sample, respond especially well to sensory descriptions. In other words, sensory descriptions and food pairing suggestions could compensate for their hesitancy to pay for unfamiliar cheeses.

“I would say the biggest takeaway of the study for the industry is to think about what type of consumer you are trying to attract and to adjust your promotional plans to match what they are looking for,” said Streletskaya, whose research broadly looks at how food labeling impacts consumer demand.

Also a factor, she said, is that sensory descriptions can be costly, depending on the retail outlet, while award stickers can be more easily incorporated in the packaging design,

Co-authors of the paper are Sara Maruyama, Susan Queisser, Sherri Cole and Juyun Lim, of Oregon State’s College of Agricultural Science, and Alina Stelick of Cornell University.

The research was supported by an OSU Dairy Foods Innovation Fund.

Cantal (IMAGE)

OREGON STATE UNIVERSITY

PrintEmail App

CAPTION

Example of image shown to consumers for Oregon State University study.

CREDIT

Oregon State University

---

JOURNAL

Food Quality and Preference

DOI

10.1016/j.foodqual.2022.104769 

ARTICLE TITLE

How information leads consumers to select specialty foods when tasting is not an option

Lots to learn from the Norwegian Public Sector's IT success

There’s been no lack of scandals in the IT industry. When NAV, the Norwegian Labour and Welfare Administration, experienced difficulties in the middle of a major project, they changed their methods – and came up with a successful solution.

Peer-Reviewed Publication

NORWEGIAN UNIVERSITY OF SCIENCE AND TECHNOLOGY

 

IMAGE: NAV, THE NORWEGIAN LABOUR AND WELFARE ADMINISTRATION, IS ON FACEBOOK, WHICH ALLOWS THE AGENCY TO QUICKLY ANSWER GENERIC QUESTIONS ABOUT BENEFITS. RECENTLY THE PUBLIC AGENCY CHANGED ITS WORKING METHOD IN THE MIDDLE OF IMPLEMENTING A NEW SOFTWARE SOLUTION FOR PAYING PARENTAL BENEFITS. NOW IT TAKES MERE SECONDS TO PROCESS AN APPLICATION FOR PARENTAL BENEFITS AND MOST CAN BE PROCESSED AUTOMATICALLY. view more 

CREDIT: THE NORWEGIAN LABOUR AND WELFARE ADMINISTRATION

Major delays, costly overruns and severe dissatisfaction among users have often characterized the introduction of new software solutions.

The industry has been learning, but there’s still room for major improvements, says Torgeir Dingsøyr, a professor at Norwegian University of Science and Technology (NTNU).

He has studied what transpired when NAV (the Norwegian Labour and Welfare Administration) changed its working method in the middle of implementing a new software solution for paying parental benefits.

From waterfall to evolution

The old Waterfall method in software development was a linear step-by-step process.

First, the project requirements were comprehensively defined, followed by major design and architecture choices, extensive testing, and finally – far down the line – delivery to the customer. Each step was dependent on the previous ones.

“The Waterfall method involves unnecessary waiting and great risk. The model’s sequential phases require the person handing off one phase to explain what happened to the next. Extensive documentation is often involved. Not an easy task. Experience shows that project staff working on different stages don't even sit together at lunch,” says Dingsøyr, who works at NTNU's Department of Computer Science.

“Even with extensive documentation, in practice the process can turn into a whispering game. The original message can be completely derailed,” he said.

Small groups

“Another well-known problem is that the IT team sits and waits for decisions in other parts of the organization. Today, the software industry has largely moved away from the Waterfall model towards the more flexible Agile methods,” says the professor.

Agile methods were developed as a response to what the industry experienced as a crisis in IT projects. Most people thought these methods only worked for small, co-located groups creating a small software product, but today they are also used in large IT projects.

NAV success

The professor thinks that most organizations today are working with what he calls first-generation large-scale agile methods for large IT projects.

When NAV started to create a computer system to simplify the processing of parental benefits in 2016, they chose a variant of the first-generation agile method.

In the middle of the project, despite a hard deadline, NAV switched over to the second-generation agile method.

The project was completed within both the deadline and budget – and to resounding success.

The processing time for applying for parental benefits was shortened from months to seconds, and 99.8 per cent of the applications were processed using self-service.

In 2019, NAV basked in the glory of being awarded the annual prize for digitalization in Norway.

Management – a support function

“In contrast to the first generation agile method, the second-generation version places the product at the centre right from the start. Work is carried out in cross-functional teams with much more freedom.”

Overlapping competence within the group means that clarification takes place continuously, instead of in large meetings that take up unnecessary time and resources.

“Less time is spent on administration and coordination, and more on product development. Management has to ensure that they set the direction for what the groups are to do – and that productivity is high. The management also becomes more of a support function for the professional work. The working groups are freed from the straitjacket of sequential phases,” Dingsøyr says.

Test immediately – learn as you go

Instead of working out a detailed and comprehensive requirement process with user participation at the beginning of a project, developers start with the most important things right away and let the users give immediate feedback, which has been found to be more effective.

“Not enough consideration has been given to user friendliness and the fact that it takes time to learn new systems. Early interaction with users can help the developers quickly understand what’s important about the system and what the customer really wants. This enables the developers to more effectively prioritize what should be included in terms of functionality,” Dingsøyr says.

“First you try a simple solution and then expand if necessary. Second-generation methods facilitate flexibility and learning along the way,” says Dingsøyr.

Effective coordinating between groups is key

Coordination is a challenge in large projects when several groups are working to create a computer system, and where there are many dependencies between the tasks.

The risk is that changes made by one person create unexpected problems for someone in another group. International studies show that work has stopped on projects where coordination between groups is problematic.

“The study of the Parental benefits project is important because it shows how the project managed to coordinate the work in a more efficient way with a second-generation method,” says Dingsøyr.

Avoid Big Bang launches!

The professor thinks it is a waste of time to wait for one big launch:

“A Big Bang launch is really risky. Technical problems can be difficult to detect before the system is deployed,” he says.

And for users, knowing what the new systems will be like is difficult before they see how the new system works in tandem with other systems they use.

It’s better to create something, test it straight away, present it – and perhaps even put it into operation before gradually launching it to more users.

The professor says that the streaming service Spotify has several groups working on functions for the app. New versions are first tested internally and then among several test users before they are released on a large scale.

“Thanks to cloud solutions, sending out new versions has become a lot easier, and different users can get different versions. We’re gaining completely different opportunities with the new technology and new methods.”

Others want to learn

Agile development methodologies are not only used by a number of large IT companies.

More and more industries are being digitized. Tesla thinks like a software company and has challenged the entire car industry. A modern car can have up to 100 million lines of software code.

“I’m seeing, for example, that Volvo Cars is adopting large-scale agile software development. Project management is showing great interest in what’s happened in the IT industry, as well as in fields such as management,” says Dingsøyr.

He believes that Norway and the other Nordic countries are leaders in software development methods. Several environments are now ripe for second-generation methods after many years of discussing software development practice.

“The study done on the Parental benefits project is the first to describe a transition to the second-generation large-scale agile development method while implementing a major IT project, and that shows how the transition led to more effective coordination practices,” says Dingsøyr.

Reference: Dingsøyr, T., Bjørnson, F.O., Schrof, J. et al. A longitudinal explanatory case study of coordination in a very large development programme: the impact of transitioning from a first- to a second-generation large-scale agile development method. Empir Software Eng 28, 1 (2023). https://doi.org/10.1007/s10664-022-10230-6

---

JOURNAL

Empirical Software Engineering

DOI

10.1007/s10664-022-10230-6 

METHOD OF RESEARCH

Case study

SUBJECT OF RESEARCH

Not applicable

ARTICLE TITLE

A longitudinal explanatory case study of coordination in a very large development programme: the impact of transitioning from a first- to a second-generation large-scale agile development method

ARTICLE PUBLICATIO