DOI: https://doi.org/10.21203/rs.3.rs-1833719/v1
Eighty two percent of in-vitro fertilization (IVF) program patients immediately ask fertility subspecialist doctors via SMS/Whatsapp messenger app when they feel different symptoms than what they usually feel. However, the high workload has been fertility subspecialist doctors’ obstacle to immediately reply these IVF program patients’ anxious questions. On the other hand, the long time it takes to receive a reply increases IVF program patients’ anxiety and, in turn, the high anxiety affects the success rate of IVF program, making it all the more important to help fertility subspecialist doctors deal with their obstacles. The weakness of Case-Based Reasoning (CBR) model is that the recommendation accuracy score is lower than that of modified CBR model and it makes the workload of fertility subspecialist doctors greater who have been busy dealing with the revision stage. To overcome this weakness, the Chris Case-Based Reasoning (CCBR) similarity formula is applied to CBR model and it combines with the Rule-Based Reasoning (RBR) model. The measurement of Confusion Matrix in the performance of CBR model combination which uses the CCBR similarity formula with RBR model suggests that the accuracy score increases to 47% and the precision score remains 100%, hence this model is declared capable of helping fertility subspecialist doctors handle the revision stage and accurately answer IVF program patients’ anxious questions.The limited scope of research makes it impossible for this research finding to be used as a standard model just yet for it needs to be tested for a wider scope.
Most in-vitro fertilization (IVF) program patients have high anxiety level[1–5] and 82% of them immediately ask questions to their fertility subspecialist doctors either via SMS or Whatsapp messenger when they have different symptoms from what they usually experience[1, 6]. IVF program patients do not care the difficulty that fertility subspecialist doctors face[7] and demand for the best services[1, 6, 8]. The high anxiety level in IVF program patients come from some factors, such as: the long period of IVF program stages they have to undergo[9, 10], the pressure to have biological children[2], the fear of failure[2, 3, 10] and the relatively expensive costs of IVF program[2, 3]. The high workload[8, 11, 12] resulting from the fact that the ratio of fertility subspecialist doctors to IVF program patients is still far from the standards set by the World Health Organization (WHO)[13, 14] or the worsening health condition[1] become obstacles for fertility subspecialist doctors to immediately reply IVF program patients’ anxious questions. As a result, the longer the time to wait for the reply from the fertility subspecialist doctor, the greater the IVF program patient’s anxiety level would be[1, 6]. On the other hand, the high level of IVF program patient’s anxiety is one factor which affects the success rate of an IVF program[1, 4, 5, 15–17]. Therefore, it is important to help lower IVF program patient’s anxiety level[4, 5, 18] by lending fertility subspecialist doctors a hand to enable them to answer all IVF program patient’s anxious questions immediately and accurately[1, 6, 7, 19–24].
A health smart system helps doctors give immediate response in handling patients[25–28] accurately[29–34]. One of the models applied in the health smart system is the Case-Based Reasoning (CBR) model[35–40]. However, this model has some weaknesses. Its first weakness is that its accuracy score is lower than that from modified CBR models, such as: Chi-square CBR (χ2 CBR) model[41] or combination of CBR model and Genetic Algorithm (GA)[42]. Its second weakness is that it gives additional workload to fertility subspecialist doctors to handle the revision stage of CBR model, when they have already had high workload[8, 11, 12], and this makes them unable to deal with the revision stage of CBR model immediately.
Mobile-based health service system (mHealth) has been developing rapidly in Indonesia. Currently, 7 mHealth applications have been provided by the government and 18 others by the private sector. The number of mHealth will keep on increasing for some factors, such as: >78% of Indonesians actively using mobile phones to access the internet, the effect of COVID-19 pandemic, doctor to patient ratio which is still far from the ideal standard (1 doctor for 3,333 patients, while WHO recommends 1 doctor for 1,000 patients), and medical personnel being not evenly distributed in every island (Indonesia has more than 13,000 islands)[14].
This research consists of 2 stages. The first stage produced a CBR similarity formula modification into Chris Case-Based Reasoning (CCBR) similarity formula applied during the Retrieve stage of CBR model. The result of first-stage research proves that applying the CCBR similarity formula in CBR model can actually improve the accuracy and precision scores of the smart system’s recommendations (in a measurable percentage). The results of first-stage research are published in Healthcare Informatics Research journal and will be the basis to conduct the second stage research, which also uses the previous research that find that the combination of CBR and Rule-Based Reasoning (RBR) model can also improve the accuracy score of a smart system’s recommendations[43–46]. For this research, the second-stage research combines CBR model that uses CCBR similarity formula and RBR model, aiming to produce a new model of smart system capable of lessening fertility subspecialist doctors’ workload in handling the CBR model’s revision stage and helping them accurately answer all IVF program patients’ anxious questions.
The stages to complete this research are:
Based on the complaint identification in IVF program patients’ records, some complaint types were obtained, such as: bleeding per vagina, suprapubic pain, fever, nausea and vomiting, and sleeping difficulty. To confirm these identification results, results from some previous research on the complaint types and treatments were sought after, as indicated in Table 1.
| Complaint Types and Their Treatments | Research Results |
|---|---|
| Bleeding Per Vagina | Results from Pontius E, et al.[49] |
| Results from Sapra KJ, et al.[50][51] | |
| Results from Sai Gnanasambanthan, et al.[52] | |
| Suprapubic Pain | Results from Sapra KJ, et al.[51] |
| Results from Sai Gnanasambanthan, et al.[52] | |
| Results from Kumar P, et al.[53] | |
| Fever | Results from Sai Gnanasambanthan, et al.[52] |
| Nausea and Vomiting | Results from Pontius E, et al.[49] |
| Results from Sapra KJ, et al.[50] | |
| Results from Sai Gnanasambanthan, et al.[52] | |
| Results from Kumar P, et al.[53] | |
| Sleeping Difficulty | Results from Goldstein CA, et al.[54] |
| Results from Huang LH, et al.[55] |
Taking the IVF program patient’s medical records, results of research on the complaint types and their treatments (information in Table 1), research results which suggests that every similar complaint for different diseases must have different weights to allow the patient’s diseases to be predicted accurately[38, 56], and research results which recommends that the weight interval should be between 0 and 1[39, 57–59] into consideration, an in-depth interview with the 2 fertility subspecialist doctors was then organized. The fertility subspecialist doctors agreed that the IVF program patients’ complaints consist of 5 complaint types, each complaint type has 4 levels, and the combination between complaint type and complaint level is referred to as complaint variation, with each complaint variation having different weight and treatment, as can be seen in Table 2. To prevent input error in choosing the complaint variation (the combination between complaint type and complaint level generates 1024 complaint variations), the computer program used in this research employs combo box facility[56].
| Complaint Type | Complaint Level | Criteria | Treatment Recommendation | Weight |
|---|---|---|---|---|
| Bleeding Per Vagina | Normal | Normal | None | 0 |
| A little bit | Blood stains | IVF program patients do not need to panic, this is normal in post ET (embryo transfer), thus only preliminary observation is needed: - Keep taking the medicine as prescribed. - Take a rest adequately. - If the complaint persists or worsens, immediately plan a visit to the Outpatient Unit | 0.1 | |
| Medium | The volume of blood secreted is around 1 table spoon | IVF program patient immediately plans for visit to Outpatient Unit | 0.5 | |
| High | Menstruation-like bleeding | IVF program patient immediately visits Emergency Unit | 1 | |
| Suprapubic Pain | Normal | Normal | None | 0 |
| Low | Still capable of doing activities normally | IVF program patients do not need to panic, this is normal in post ET (embryo transfer) or it is a sign of nidation (< 1 week post ET), it is recommended to have preliminary observation: - Keep taking the medicine as prescribed. - Stay hydrated. - If the complaint persists or worsens, immediately plan a visit to the Outpatient Unit | 0.1 | |
| Medium | Complaints disrupt activities | IVF program patient immediately takes paracetamol and plans for visit to Outpatient Unit | 0.5 | |
| High | Unable to do any activity | IVF program patient needs to immediately have themselves checked at Emergency Unit | 1 | |
| Fever | Normal | Body temperature <37.2°C | None | 0 |
| Low | Body temperature 37.2°C − 37.5°C | IVF program patient needs to make preliminary observation: - Keep taking the medicine as prescribed. - Stay hydrated. - If the complaint persists or worsens, immediately plan a visit to the Outpatient Unit | 0.1 | |
| Medium | Body temperature 37.5°C − 40°C | IVF program patient immediately takes paracetamol and plans for visit to Outpatient Unit | 0.5 | |
| High | Body temperature >40°C | IVF program patient immediately visit Emergency Unit | 1 | |
| Nausea and Vomiting | Normal | Normal | None | 0 |
| Low | Capable of doing activities normally | IVF program patient needs to make preliminary observation: - Keep taking the medicine as prescribed. - Measure and record your waistline. - Keep calm and reduce your anxiety - If the complaint persists or worsens, immediately plan a visit to the Outpatient Unit | 0.11 | |
| Medium | Activities are disrupted | IVF program patient immediately visits Emergency Unit | 0.5 | |
| High | Unable to do any activity | IVF program patient immediately visits Emergency Unit | 1 | |
| Sleeping Difficulty | Normal | > 7 hours | None | 0 |
| Low | 6–7 hours | IVF program patient needs to make observation: - Keep taking the medicine as prescribed. - Eat adequately and stay hydrated - Keep on doing activities normally - Keep calm and reduce your anxiety - If the complaint persists or worsens, immediately plan a visit to the Outpatient Unit | 0.25 | |
| Medium | 5–6 hours | IVF program patient immediately plans for visit to Outpatient Unit | 0.5 | |
| High | < 5 hours | IVF program patient immediately visits Emergency Unit | 1 |
Based on information in Table 2, complaint variations as in Table 3 can be identified and their treatments as in Table 4 can be recommended. Furthermore, a Rule-Based Reasoning (RBR) for an IVF program patient’s anxious question replier smart system is prepared as in Fig. 2.
| Code | Complaint Type | Complaint Level |
|---|---|---|
| K1-1 | Complaint type: bleeding per vagina | Complaint level: mild or low |
| K1-2 | Complaint level: medium | |
| K1-3 | Complaint level: heavy or high | |
| K2-1 | Complaint type: suprapubic pain | Complaint level: mild or low |
| K2-2 | Complaint level: medium | |
| K2-3 | Complaint level: heavy or high | |
| K3-1 | Complaint type: fever | Complaint level: mild or low |
| K3-2 | Complaint level: medium | |
| K3-3 | Complaint level: heavy or high | |
| K4-1 | Complaint type: nausea and vomiting | Complaint level: mild or low |
| K4-2 | Complaint level: medium | |
| K4-3 | Complaint level: heavy or high | |
| K5-1 | Complaint type: sleeping difficulty | Complaint level: mild or low |
| K5-2 | Complaint level: medium | |
| K5-3 | Complaint level: heavy or high |
| Code | Complaint Treatment Recommendation |
|---|---|
| R-1 | IVF program patient needs to perform observation for bleeding per vagina complaint |
| R-2 | IVF program patient needs to perform observation for suprapubic pain complaint |
| R-3 | IVF program patient needs to perform observation for fever complaint |
| R-4 | IVF program patient needs to perform observation for nausea and vomiting complaint |
| R-5 | IVF program patient needs to perform observation for sleeping difficulty complaint |
| R-6 | IVF program patient immediately plans to have themselves examined to Outpatient Unit |
| R-7 | IVF program patient immediately takes paracetamol and plans to have themselves examined to Outpatient Unit |
| R-8 | IVF program patient immediately has themselves examined at Emergency Unit |
| R-9 | No recommendation for IVF program patient is available |
To determine the accuracy level of recommendations, a standard score for reference is needed. If the score is below this standard, the recommendation is declared inaccurate and if it is at least equal with the standard score, the recommendation is declared accurate. CBR model[38] has not capable of setting a minimum score standard to determine whether or not the recommendations can be declared accurate that it can be directly sent to the IVF program patient or declared inaccurate that it needs to be validated first by the fertility subspecialist doctors. Some previous research recommends to set the standard score at 80% as can be seen in Table 5.
| Research Results | Recommendation |
|---|---|
| Elisabeth Beaunoyera, et al. | Score ≥ 80%, the information produced has higher quality and is reliable [60]. |
| Marisa Louridas, et al. | The minimum score agreed upon by experts is 80%[61]. |
| Justin Parent, et al. | The minimum reliability score is 80%[62]. |
Based on information in Table 5, this research sets the minimum standard score at 80% in CBR model which uses CBR similarity formula[1] or which uses CCBR similarity formula, as seen in Fig. 3. The 80% minimum standard score is also applied to the combination between CBR model (using CBR similarity formula atau CCBR similarity formula) and RBR model (Fig. 1), as seen in Fig. 4. Table 6 is the test model in this research which was tested using the same test data.
Figure 4. The combination of CBR model (CBR similarity formula or CCBR similarity formula) and RBR model
The second-stage research uses 2 similarity formulas applied to CBR model, namely CBR similarity formula[1, 6] and CCBR similarity formula (results of first-stage research). The differences between the two similarity formulas are explained in Fig. 5. CCBR similarity formula adopts Dekang Lin’s[63, 64] similarity theory which suggests that to determine the similarity between A and B, the similar and different factors need to be considered. The more factors that A and B have in common, the more similar they are, the more different factors that A and B have, the more dissimilar they are. To discover the similarity and difference between A and B, they need to be viewed from different perspectives and each perspective should calculate the similarity level. The optimal similarity score is the mean of the similarity score of each perspective.
| Test Model | Similarity Formula | Logical Flowchart |
|---|---|---|
| CBR model | CBR similarity formula | Logical flowchart in Fig. 3 |
| CBR model | CCBR similarity formula | Logical flowchart in Fig. 3 |
| Combination between CBR model and RBR model | CBR similarity formula | Logical flowchart in Fig. 4 |
| Combination between CBR model and RBR model | CCBR similarity formula | Logical flowchart in Fig. 4 |
Out of 97 recommendations produced by CBR model which uses the logical flowchart in Fig. 3, only 92% of the recommendations generated by the CBR model which uses CBR similarity formula have similarity score ≥ 80%, and only 42% generated by the CBR model which uses CCBR similarity formula, as shown in Fig. 5. Furthermore, the model’s recommendations were validated by the 2 fertility subspecialist doctors to ensure that they match the fertility subspecialist doctors’ recommendations, as shown in Fig. 6.
The two fertility subspecialist doctors agreed that 92% of recommendations generated by the CBR model which uses CBR similarity formula had similarity score ≥ 80%, yet only 27% of these recommendations were declared valid, and 42% of the recommendations proposed by the CBR model which uses CCBR similarity formula had similarity score ≥ 80%, yet 100% of them were declared valid, as shown in in Fig. 7. The validation of fertility subspecialist doctors proves that not every one of the model’s recommendations with similarity score ≥ 80% match those that fertility subspecialist doctors would make, only recommendations generated by the CBR model which uses CCBR similarity formula are trustworthy. The excellent performance that the CBR model which uses CCBR similarity formula achieves is because of the performance of CCBR similarity formula which adopts Dekang Lin’s[63] similarity theory, making it more precise in calculating the similarity score.
The validation from the fertility subspecialist doctors proves that CBR model which uses CCBR similarity formula has the best performance and information in Fig. 8 is the result of performance measurement using Confusion Matrix.
The excellent performance that the CBR model which uses CCBR similarity formula achieves is also shown in the result of validation by the fertility subspecialist doctors who suggests that out of the recommendations with similarity score < 80% (information in Fig. 5), 82% of them are declared valid, as shown in Fig. 9. However, any recommendation with similarity score < 80% cannot be used since this research finds that the highest similarity score for invalid result from the CBR model which uses CCBR similarity formula is 79.7%.
In the CBR model, to deal with the recommendations with similarity score < 80%, the fertility subspecialist doctors need to go to revision stage to validate the recommendations. The reality is that fertility subspecialist doctors have already had high workload[1, 8, 11, 12], and this makes them unable to immediately revise the CBR model’s recommendations. On the other hand, if IVF program patients wait for the reply for too long, it might increase their anxiety level[1, 6] and adversely affect the IVF program they are participating in[1, 4, 5, 15–17]. To handle this, the CBR model was modified by combining CBR model and RBR model[43–46], following the logical flowchart in Fig. 4. The two fertility subspecialist doctors were still involved to validate the model’s recommendations and the performance was measured using Confusion Matrix, as shown in Fig. 10.
The result of Confusion Matrix measurement proves that the combination between CBR model and RBR model can improve the CBR model’s performance. The RBR model’s success to improve the CBR model’s performance is because the Rule-Based Reasoning properly adopts the way the 2 experienced fertility subspecialist doctors work in dealing with IVF program patient’s complaints and uses CCBR similarity formula. The highest improvement of accuracy score is shown by the combination of CBR model which uses CCBR similarity formula and RBR model at 47% with the precision score staying at 100% (Fig. 8) or without any increase (0%). The improved accuracy score and the high precision score proves that the combination between CBR model which uses CCBR similarity formula and RBR model can actually help the fertility subspecialist doctors deal with the revision stage of CBR model and accurately reply the IVF program patient’s anxious questions.
Considering some factors, such as: >78% Indonesians actively using mobile phone to access the Internet, fertility subspecialist doctor-IVF program patient ratio still below WHO’s standard[14] and the excellent performance of the combination between CBR model (using CCBR formula) and RBR model, this research recommends an architecture of IVF program patient’s anxious question replier smart system as shown in Fig. 11. Based on the logical flowchart in Fig. 2, an algorithm of combination between CBR model (using CCBR similarity formula) and RBR model which complements the information of architecture of IVF program patient’s anxious question replier smart system as shown in Fig. 12 is built.
The logical model applied to build the smart system software is the combination between CBR model (using CCBR similarity formula and applying the standard score of ≥ 80% for recommendation accuracy level) and RBR model. The smart system software model uses the hybrid model (web and mobile applications), hence the smart system users (fertility subspecialist doctors, IVF program patients, and hospital medical staff) can access the smart system software via the Internet network connected to many devices they use, such as computer, notebook or smartphone. To build the database and web technologies in the smart system, it is recommended to use opensource software which has some strengths, such as: being used by many programmers, having high reliability and good quality[65].
We recommend the combination between CBR model which uses CCBR similarity formula and applies the standard score ≥ 80% for its recommendation accuracy level and RBR model to be the right model to be applied to the IVF program patient’s anxious question replier smart system. This combination model has been found capable of helping fertility subspecialist doctors deal with the CBR model’s revision stage and supporting fertility subspecialist doctors to keep on providing high-quality health services[66], by providing accurate recommendations to answer IVF program patient’s anxious questions. For the smart system software model to be built, it is recommended to use the hybrid model to allow the smart system users to immediately access the smart system via many devices they own.
The fact that this research only involved 2 fertility subspecialist doctors and 2 hospitals providing the IVF program service in Central Java and Yogyakarta Provinces prevents this research result from being eligible to be a standard model yet since it still needs further test for a wider scope and this is certainly an opportunity for further research.
Acknowledgments
The writer would like to thank Dr. Doddy Sutanto, M.Kes, SpOG(K)Fer who has provided valuable feedbacks in this research.
Ethical Approval
The Ethical Committee of Medicine Committee at Duta Wacana Christian University under a letter No. 1384/C.l6/FK/2022
Competing interests
No conflict of interest is involved in the publication of this research result.
Authors' contributions
Dr. Irwan Sembiring as the first writer: do the analysis and validation of experimental results
Paminto Agung Christianto as the second author: generate ideas, collect data and conduct experiments
Prof. Dr. Eko Sediyono as the third author: do the analysis and validation of experimental results
Funding
Funding for research activities comes from the researcher's personal funds
Availability of data and materials
Considering the “Ministerial Regulation of Health of the Republic of Indonesia No. 269/MENKES/PER/III/2008” into account, the medical records obtained and used are those containing complaints and recommendations for treating the complaints without any personal information of the IVF program patients (experimental data attached).