This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Pollen forecasts are highly valuable for allergen avoidance and thus raising the quality of life of persons concerned by pollen allergies. They are considered as valuable free services for the public. Careful scientific evaluation of pollen forecasts in terms of accurateness and reliability has not been available till date.
The aim of this study was to analyze 9 mobile apps, which deliver pollen information and pollen forecasts, with a focus on their accurateness regarding the prediction of the pollen load in the grass pollen season 2016 to assess their usefulness for pollen allergy sufferers.
The following number of apps was evaluated for each location: 3 apps for Vienna (Austria), 4 apps for Berlin (Germany), and 1 app each for Basel (Switzerland) and London (United Kingdom). All mobile apps were freely available. Today’s grass pollen forecast was compared throughout the defined grass pollen season at each respective location with measured grass pollen concentrations. Hit rates were calculated for the exact performance and for a tolerance in a range of ±2 and ±4 pollen per cubic meter.
In general, for most apps, hit rates score around 50% (6 apps). It was found that 1 app showed better results, whereas 3 apps performed less well. Hit rates increased when calculated with tolerances for most apps. In contrast, the forecast for the “readiness to flower” for grasses was performed at a sufficiently accurate level, although only two apps provided such a forecast. The last of those forecasts coincided with the first moderate grass pollen load on the predicted day or 3 days after and performed even from about a month before well within the range of 3 days. Advertisement was present in 3 of the 9 analyzed apps, whereas an imprint mentioning institutions with experience in pollen forecasting was present in only three other apps.
The quality of pollen forecasts is in need of improvement, and quality control for pollen forecasts is recommended to avoid potential harm to pollen allergy sufferers due to inadequate forecasts. The inclusion of information on reliability of provided forecasts and a similar handling regarding probabilistic weather forecasts should be considered.
Pollen allergies are a global health concern [
Data quality is a concern [
A range of mobile apps relevant to pollen allergies which provide pollen forecasts are freely available, but users have not been informed about the accurateness of pollen forecasts till date. Simultaneously, the identity of the publishers of such apps with health-related content for pollen allergy sufferers has not been available. Also, studies on apps with a focus on pollen allergies and allergen prevention are still rare [
Herein, we analyze the grass pollen forecasts of 9 mobile apps from four different countries during the grass pollen season of 2016. Moreover, the forecast dates for the readiness to flower for grasses in order to examine the accurateness and reliability of these forecasts was evaluated. Gaps and errors in the forecasts are also considered in order to complete the picture of the performance of specific apps.
The following free downloadable apps were included in this study: (1) “Pollen” [
The pollen data used was retrieved from the European Aeroallergen Network (EAN) database and originates from the national pollen monitoring networks (see “Acknowledgments”). Pollen data was attained by Hirst-type pollen traps [
Descriptive statistics and R3.3.1 (R Foundation for Statistical Computing, Vienna, Austria) were used for the analysis. Daily grass pollen concentrations were compared within this defined grass pollen season with the forecasted loads. Used load classes and their respective pollen concentration ranges are shown in
In addition, the performance of the forecast “readiness to flower” for grasses was compared with the first-occurring moderate pollen load in the season. The first moderate grass pollen load was chosen as “threshold,” as low grass pollen loads can occur in the air preceding the grass pollen season. However, only two pollen apps (“Pollen” and “Pollen News”) provided such a service.
The mobile apps under study herein are characterized concerning their main options, functional background, and noteworthy observations. All of them are available for free.
“Pollen” comprises pollen information (forecasts, countdown for the season, different pollen dispersal models, and forecast maps), services concerning pollen allergy (diary for allergic symptoms and burden, daily load based on symptom entries, personalized pollen information, search for allergologists), botanical information, push notification services, and an imprint (the Austrian pollen information service is the developer and operator). To the best of our knowledge, it is the only app described and published scientifically [
“Biowetter” is an app focusing on weather and biorhythm, thus pollen information is only one part of the app. An imprint does not exist and advertisement appears regularly in the app. During the study period, it was the app with the most frequent crash reports and missing forecast information (
“Pollenwarner” includes pollen information, push notification services, a symptom diary, and general information (tips and tricks). An imprint is not available, but viewing the logos indicates that the companies Tempo and Otriven are connected with this app.
The forecasted load classification is shown along with the respective pollen concentrations used in this study. The intermediate stages (no-low, low-moderate, moderate-high) are used only by one app (“DWD”).
Forecast load | Daily pollen concentrations |
No | 0-0.99 |
No-low | 0-2 |
Low | 1-19.9 |
Low-moderate | 15.6-23.4 |
Moderate | 20-49.9 |
Moderate-high | 39.6-59.4 |
High | 50 and above |
The results of the hit rates are shown per app and per analyses (ie, exact hit rates and hit rates with tolerances of ±2 or ±4 pollen, respectively). Note the increasing hit rates in some apps (such as “Pollen,” “Allergiehelfer,” and “Pollenflug”) versus the less improving apps (such as “Hayfever”).
Mobile apps | Exact hit rates |
Hit rates with tolerance 2 |
Hit rates with tolerance 4 |
“Pollen” | 62.9 | 77.3 | 80.4 |
“Biowetter” | 31.8 | 39.4 | 42.2 |
“Pollenwarner” | 34.0 | 39.2 | 42.3 |
“DWD” | 41.1 | 46.6 | 52.1 |
“Allergiehelfer” | 48.6 | 54.3 | 57.1 |
“Pollenflug” | 50.7 | 56.2 | 58.9 |
“Allergohelp” | 45.2 | 50.7 | 52.1 |
“Pollen News” | 42.4 | 48.5 | 51.5 |
“Hayfever” | 35.7 | 39.3 | 39.3 |
“DWD” provides pollen forecasts, a forecast map, and possesses an imprint, “Deutscher Wetterdienst,” which is not only the official German weather service, but also a higher federal authority in the business area of the Federal Ministry of Transport and Digital Infrastructure that provides pollen forecasts based on the pollen measurements from the Foundation German Pollen Information (PID).
“Allergiehelfer” comprises pollen forecasts including push-service and general information. The pharmaceutical company GlaxoSmithKline GmbH & Co. KG is indicated in the imprint.
“Pollenflug” delivers pollen forecasts, a forecast map, push notification services, an allergy questionnaire, and general information. The pharmaceutical company Hexal AG is mentioned in the imprint and the app includes advertisement of allergy medication.
“Allergohelp” provides pollen forecasts, forecast maps, general information on allergy and therapy, therapy documentation, and a search for allergologists. The pharmaceutical company Allergopharma GmbH & Co. KG is mentioned in the imprint.
“Pollen-News” comprises pollen information (pollen forecasts, flowering start, maps) and general information (tips and tricks, information on aeroallergens). The aha! Allergiezentrum Schweiz and Bundesamt für Meteorologie und Klimatologie MeteoSchweiz are indicated as publisher in the imprint. The aha! Allergiezentrum Schweiz sees itself as patient organization, whereas the Bundesamt für Meteorologie und Klimatologie of MeteoSchweiz is the national weather and climate service in Switzerland.
“Hayfever” provides pollen forecasts and forecast maps. Advertisement is present and an imprint does not exist. However, the company A.Vogel (natural and herbal remedies) is present in the form of a logo.
The quality of the grass pollen forecast, measured by hit rates (
Boxplots of the forecasted pollen level for all apps analyzed revealed further insights (
Grasses were ready to flower based on the forecast countdown of “Pollen” in Vienna on May 9, 2016; and the first moderate grass pollen load was also occurred on the very same day. Grasses were ready to flower based on the forecast countdown of “Pollen News” in Basel on May 10, 2016; and the first moderate grass pollen load occurred 3 days earlier (ie, on May 7, 2016).
Forecasting pollen concentrations or loads is not a trivial concern since many factors play a role in the development of the burden for pollen allergy sufferers. Several data sources and information are required including the biogeography of the region, continuous pollen monitoring and reliable pollen data, high quality weather forecasts, phenology, models (numerical simulations of pollen dispersal), medical-allergological expertise respective symptom data, and experience in the task of pollen forecasting [
Although only two apps provided a forecast on the readiness to flower (“Pollen” and “Pollen News”), those forecasts had an excellent performance. The forecasted dates used here were the ones used when the information “ready to flower” occurred. Earlier forecast dates from mid-April (eg, April 15, 2016) announced for both apps May 6, 2016 as date for grasses ready to flower, which is still accurate enough considering it was a forecast from nearly a month before the start of the respective grass pollen season. The forecasted date coincided with the first moderate grass pollen load (“Pollen”) or within a few days (“Pollen News”) and may thus be recommended as a useful service for pollen allergy sufferers to prepare them for the grass pollen season.
Boxplots of the apps under study and their performance concerning forecasting grass pollen loads (none, low, moderate, high) and missing forecasts (NA).
Certain facts and dependencies complicating the generation of an accurate pollen forecast have to be mentioned. Pollen forecasting activities depend on weather forecasts that are as accurate as possible. Although weather forecasting accuracy hit rates range from 85% to 95% for minimum and maximum temperatures on the same day, the forecasts of precipitation and sun hours are more complex and achieve lower rates [
Mobile apps dealing with allergen avoidance with the aim of supporting pollen allergy sufferers should fulfill certain criteria and functionalities, among them easily understandable pollen forecasts, a minimum of forecasted aeroallergens, botanical information, symptom diaries, allergy risk questionnaires, and an imprint with the publisher of the app stating the responsible institution, at best without conflict of interests [
A combination of data sources and methods will lead to an improvement of pollen forecasts and is already used in the mobile app with the best performance found in this study. Phenological routines assess the local progression of the pollen season. Pollen dispersal models and readiness to flower models support the person preparing the pollen forecast. Symptom data reveal the impact of the daily pollen concentrations on pollen allergy sufferers and allows for tailoring the forecast to the needs of persons concerned. These are possible gateways in the need for a tighter connection to a future pollen forecast. Specific recommendations concerning the improvement of pollen forecasts comprise (1) ongoing evaluation of pollen forecast quality at best also during the pollen season to raise quality as fast as possible, (2) comprehension of the probabilistic nature of pollen forecasts and implementation of this aspect in services visible for users (such as percentual information on the forecast accuracy), and (3) implementation of all information sources necessary such as symptom data and phenological routines besides pollen data to improve a pollen forecast [
Pollen forecasts are essential for pollen allergy sufferers in terms of allergen avoidance and thus the accuracy of such forecasts is a key factor for improving the quality of life. Most apps deliver forecasts with a hit rate of about 50%, which is a score that is too low for this purpose. Quality control of pollen forecasts should be introduced since wrong forecasts can be seen as potential physical injury and may harm persons concerned significantly. Pollen information and pollen forecasts should never be given out by pharmaceutical companies or be accompanied by advertisement [
European Aeroallergen Network
We thank the following institutions and persons for supplying pollen data: MeteoSwiss (for Basel), Foundation German Pollen Information (PID), Prof Karl-Christian Bergmann, Matthias Werchan (Berlin), and Met Office in Exeter (London). Pollen data from Vienna originates from the Austrian pollen information service and was supplied by one of the authors (MK). We are grateful to Alexander Kowarik for support in the statistical analyses. Gina Semprebon (Bay Path University) proofread the manuscript. It is our wish to express also our gratitude to two anonymous reviewers, who contributed with their insights to the manuscript.
The study was designed by all authors (KB, UB, and MK). Data analyses were performed by KB and MK. Technical and scientific supervision was assured by UB. All authors were involved in data interpretation, drafting the manuscript, editing, and final approval.
KB, MK, and UB report to have taken part in the development of the app “Pollen” that is freely available and without advertisement and thus is of no financial interest.