Reliability of Results from Randomized Studies and Systematic Reviews in Acupuncture: A Methodological Review for Acupuncture Against Symptoms of Breast Cancer Care.

Introduction
Acupuncture is a method of treatment used in traditional Chinese medicine (TCM) for many different diseases and ailments. In the Western world it has gradually been adapted to deal with a wide range of different disorders including pain and migraines and other conditions (1, 2, 3). Even though the mechanism behind acupuncture is unclear, acceptance of this therapy is widespread in Western health care systems. Both lay persons as well as health care professionals have been won over to its efficacy. Acupuncture has also been used as a complementary treatment that is combined with conventional cancer therapy to reduce different types of side effects (4). One of these fields of application are menopausal symptoms in women with breast cancer (4, 5).
In hormone-receptor-positive breast cancer, antihormonal therapy with aromatase inhibitors or tamoxifen is an important part of the therapy strategy (6, 7). However, menopausal symptoms are a side effect of this therapy. Menopausal symptoms include many different complaints, such as hot flashes and musculoskeletal symptoms (8, 9). These problems can be very stressful for the patients and they greatly reduce compliance with the therapy (7). Thus, strategies have been sought to manage these side effects; different drugs are also available as, for example, venlafaxine in the case of hot flashes (10, 11). Moreover, many patients do not want more medication (12). Therefore, alternative treatments from complementary medicine have been sought out. Acupuncture has fewer adverse effects than some drugs (13). Many studies have been conducted in search of evidence that shows an effect of acupuncture. However, it remains unclear whether acupuncture has a specific or unspecific effect.
Despite the unclear study results, the official guidelines in Germany for treating breast cancer contain recommendations in favor of acupuncture, for example in the guideline on breast cancer regarding musculoskeletal complaints (14). Acupuncture is also recommended in the guideline on complementary medicine, for example for hot flashes (5). Not only in Germany, but also internationally, Greenlee’s guidelines on integrative therapy for breast cancer, for example, take a positive view of acupuncture and recommend it as an option for hot flashes (4).
In order to provide the best possible care for patients with breast cancer, it is important to know which strategies are effective for managing side effects. In addition, methodologically flawed studies may lead to incorrect conclusions in practice. Therefore, we decided to conduct a systematic review on the existing literature on acupuncture used to manage menopausal symptoms (hot flashes and musculoskeletal pain) in breast cancer patients receiving endocrine treatment.
It should be noted, however, that the mechanism behind the effect of acupuncture is still unclear. The traditional explanation maintains that acupuncture brings life energy, called Qi, back into balance. This happens by manipulating certain points, which are located on so-called meridians, with acupuncture needles (15-17). There are modern theories regarding the potential mechanisms of action of acupuncture. For example, acupuncture may influence the extracellular matrix and tissue mechanobiology by activating mechanosensitive ion channels and integrin signaling (18). Additionally, acupuncture may impact the body’s biotensegrity system through fascial networks, influencing mechanotransduction across multiple tissue levels (19, 20). It has not yet been possible to conclusively clarify how an effect can be achieved (21).

Materials and Methods
For a comprehensive overview, first of, RCTs were searched and the results of the studies were analyzed. The studies were then assessed using the Risk of Bias 2 (RoB) tool and the results were critically appraised. In addition to the RoB tool, other criteria that were not covered by the RoB tool were included in the evaluation. These critiques are specific to acupuncture and are intended to show the complexity of this intervention and its influence on the validity of the effects of acupuncture. Next, SRs and Meta analyses (MAs) were searched, reviewed and the results examined in conjunction with the results of the individual studies. In addition, the methodological evaluation of the SRs in the individual studies was compared with each other and with our own evaluation to assess differences.
Search strategy and study selection. Independent systematic searches were conducted for RCTs and SRs/ MAs in five databases in English language [Medline (Ovid), Embase (Ovid), Cochrane, PsycINFO (EBSCO), CINAHL (EBSCO)]. In February 2025 a complex search strategy was developed for each database consisting of Mesh Terms, keywords, and text words with different spellings (see Table S1 and Table S2). They were related to menopausal symptoms, anti-hormonal therapy, acupuncture therapy and breast cancer. The search results were imported into EndNote and, after identifying and removing all duplicates, a title abstract screening was conducted by one reviewer (KLK) and controlled by another reviewer (JD). In case of disagreement consensus was reached by discussion. A full-text copy was retrieved if the title and abstract did not provide sufficient information for screening. Thereafter full texts were retrieved and screened and checked again by two reviewers. Thereafter full texts were retrieved and screened and checked again by two reviewers (KLK, JD).
Inclusion and exclusion criteria. The inclusion and exclusion criteria are based on a PICO model and are listed in Table I and Table II.
Assessment of risk of bias. The risk of bias of the included RCTs was assessed using the Revised Cochrane risk-of-bias tool for randomized trials (RoB2). Characteristics were assessed by one reviewer (KLK) and controlled by another reviewer (JH/JD). In case of disagreement consensus was made by discussion. If there were more than two groups, the RoB assessment was performed for the acupuncture group and the active comparison group.
The RoB for the SRs was analyzed using the AMSTAR-2 instrument. Characteristics were assessed by one reviewer (KLK) and controlled by another reviewer (JD). In case of disagreement consensus was made by discussion. Shea et al. (22) identified seven critical items in the AMSTAR-2 tool to determine the overall confidence in the results of systematic reviews (SRs). However, we diverged from their approach by rating item 8 (providing a detailed description of the included studies) as a critical item instead of item 7 (providing a list of excluded studies). Additionally, we excluded item 2 (reporting an explicit statement that review methods were established prior to conducting the review and justifying deviations from the protocol) as a critical item. These changes were made in alignment with the specific requirements of our research question.

Results
Characteristics of included studies (RCTs). The systematic search found 628 articles, eight were added by manual searching. Of these, 124 were identified as duplicates, which led to a title abstract screening of 512. A total of 462 studies could be excluded with the title abstract screening, leaving 50 for the full text screening. Finally, 22 studies were included. The other studies were excluded due to study type and publication type (e.g., conference proceeding) (PRSIMA diagram, Figure 1; for excluded studies see Table S4).
The 22 studies included were published between 2005 and 2024. The method of acupuncture differs in the studies; traditional acupuncture was used in 15 and electroacupuncture in 7. In addition, different points were chosen for acupuncture. Moreover, the studies differed in the number of groups: 17 studies had two groups, 4 had three groups, and one study had four groups. The groups were also designed differently: in 14 trials acupuncture was compared with sham acupuncture, in 6 trials there was no treatment/waitlist/delayed acupuncture (23-28), in two trials hormone therapy was used for comparison (29, 30), in two trials “applied relaxation” was the comparison (31, 32), in one trial “enhanced selfcare” was used as the comparison (33). One trial used gabapentin and a gabapentin placebo (34). Another trial used venlafaxine (35).
Characteristics of SRs included. The systematic search revealed 108 articles; two additional articles were added by manual search. First, duplicates were removed, leaving 99 articles. A title-abstract screening was then carried out, leaving 25 publications for the full-text screening. Finally, 16 SRs or MAs were included, excluding nine publications, as they did not meet the inclusion criteria (PRISMA diagram Figure 2, for excluded studies see Table S5). Of these 16 articles, six articles did not only analyze the effect of acupuncture on menopausal symptoms, but they also assessed other methods against menopausal symptoms, such as prescription drugs and yoga (36-41). For this analysis, only the data on acupuncture interventions and the control group were considered. The majority of the papers focused on only two menopausal symptoms: hot flashes [seven studies (36, 39, 42-46)] and arthralgia/musculoskeletal symptoms [six studies (37, 38, 40, 41, 47, 48)]; three other studies included both (49-51). Eleven studies conducted a meta-analysis as well (37, 41-49, 51). Of these, seven did this for hot flashes (42-46, 49, 51) and only four did it for musculoskeletal symptoms (37, 41, 47, 48). The SR of Pan et al. (51) is an update meta-analysis of the SR of the same working group [Pan et al. (49)]. Table S3 provides an overview of the studies included in the SRs.
Methodological quality assessment. RoB of RCTs (RoB Tool. 2.0). In the RoB assessment using the RoB-2 instrument, all 22 studies were rated as having ‘high concerns’, with none rated as having ‘some concerns’ or ‘low concerns’ (Figure 3). Common methodological problems were missing outcome data, insufficient information on blinding (e.g., insufficient information about allocation sequence concealment), and inadequate data analysis or inadequate information about data analysis. It is important to consider the methodological assessment when interpreting the results. As studies with high concerns must be viewed and interpreted with caution, only a brief outline of the results of these included.
Risk of bias of systematic reviews (AMSTAR-2). In the AMSTAR assessment using the AMSTAR 2, none of the 16 SRs received a rating of high overall confidence or moderate overall confidence. Six SRs were rated low overall confidence, and 10 SRs were rated critically low overall confidence (Figure 4). Common methodological problems were not specifying the review methods prior to the review, not explaining the choice of study type, not providing a list of excluded studies, not adequately describing the included studies, not considering the RoB when interpreting results, and not considering publication bias in the quantitative synthesis.
Results of the RCTs and SRs. Hot flashes (RCTs). Fifteen of the included RCTS deal with hot flashes (23, 27-35, 52-56). All RCTs were rated with high concerns.
Seven studies compared RA or EA to SA (23, 34, 52-56). Of these, five had two groups comparing RA and SA. Two studies found no significant difference between RA and SA (52, 54), whereas three studies found RA to be significantly better than SA (53, 55, 56). One study compared RA to SA to a passive control group and found RA to be significantly better than both SA and the passive control group (23). One study compared EA to SA, gabapentin and a placebo pill, and found that SA and EA were significantly better than the placebo pill (34) (Table III).
Hot flashes (SRs). Ten SRs on hot flashes were included (36, 39, 42-46, 49-51), seven of which performed a meta-analysis (42-46, 49, 51). Six of these SRs (2 low, 4 critically low overall confidence) assessed the reduction/number of hot flashes (42-44, 49-51), five (5 critically low overall confidence) the Kupperman index or menopausal symptoms (43-45, 49, 51), one (critically low overall confidence) a hot flash score (46) and six (2 low, 4 critically low overall confidence) the intensity and frequency of hot flashes (36, 39, 43-46). Of the meta-analyses, three (3 critically low overall confidence) calculated the number/reduction of hot flashes (43, 49, 51), five (5 critically low overall confidence) the Kupperman index or menopausal symptoms (43-45, 49, 51), four (1 low, 3 critically low overall confidence) the frequency of hot flashes (42, 44-46), one (critically low overall confidence) a hot flash score (46) and one (critically low overall confidence) the intensity of hot flashes (45).
Seven (36, 43-46, 49, 51) of the SRs received rating critically low overall confidence in the AMSTAR assessment and 3 (39, 42, 50) received the rating low overall confidence. There are indications in the SRs with a low overall confidence rating that acupuncture could have an effect on hot flashes (42, 50). However, due to methodological problems, these cannot be confirmed with certainty (39) (Table IV).
Musculoskeletal symptoms (RCTs). Seven RCTs focused on musculoskeletal symptoms (24-26, 57-60). All seven studies have a high risk of bias rating.
Four studies compared RA or EA to SA. Three studies (57-59) had two groups comparing RA and SA, of which only 1 study found RA to be significantly better than SA (59). One study compared EA with SA and found no significant differences between groups (60). Three studies compared RA to SA and to a passive control group. Three studies had three groups comparing RA and SA to WLC, all found RA to be significantly better than controls (24-26) (Table V).
Musculoskeletal symptoms (SRs). Nine SRs on musculoskeletal symptoms were included (37, 38, 40, 41, 47-51), 6 of which calculated an MA (37, 41, 47, -49, 51). Eight of these studies (4 low, 4 critically low overall confidence) analyzed musculoskeletal pain (37, 38, 40, 41, 48-51), two (2 critically low overall confidence) stiffness (49, 51), two (2 critically low overall confidence) Brief Pain Inventory (BPI) (37, 47) and one (critically low overall confidence) Western Ontario and McMasters Universities’ Osteoarthritis index (WOMAC) (47). Meta analyses were calculated from four studies (1 low, 3 critically low overall confidence) on pain (41, 48, 49, 51), two (2 critically low overall confidence) on stiffness (49, 51), two (2 critically low overall confidence) on BPI (37, 47), and one (critically low overall confidence) on WOMAC (47).
Five (37, 41, 47, 49, 51) of the SRs received rating critically low overall confidence in the AMSTAR assessment and four (38, 40, 48, 50) received the rating low overall confidence. The four SRs with the rating low overall confidence cannot find any reliable evidence for the effectiveness of acupuncture (38, 40, 48, 50) (see Table VI).
Further risk of bias assessments in acupuncture studies. All included studies show high risk in the evaluation with the Cochrane RoB tool 2.0. Still, considering studies on acupuncture, the RoB-2 instrument may not include on all aspects explicitly to be considered when evaluating the reliability of a study. For example, there are some aspects which could lead to unblinding of the sham and the verum arm in the course of the study. This is especially important to consider in studies in which RA is compared to SA. The following aspects might lead to unblinding or to some notion of the patient of which arm he/she is:
Different number of points in RA and SA: Bao et al. (57, 58), Oh et al. (60), Serra et al. (56); Different penetration depth in RA and SA: Hervik et al. (53, 55), Crew et al. (59), Bokmand et al. (23); Hershman et al. (24, 26); Different localization of the points: Bao et al. (57, 58), Hervik et al. (53, 55), Deng et al. (52), Crew et al. (59), Liljegren et al. (54), Bokmand et al. (23), Mao et al. (25, 34), Hershman et al. (24, 26), Serra et al. (56), Oh et al. (60); Individual acupuncture points in RA vs. standardized protocol in SA: Crew et al. (59), Mao et al. (25, 34);
Non-penetrating needles – sometimes fixed with adhesive tape (as SA): Deng et al. (52), Liljegren et al. (54), Bao et al. (57, 58), Oh et al. (60), Mao et al. (25, 34), Serra et al. (56); In case of EA: current applied in RA vs. no current in SA: Oh et al. (60), Mao et al. (25, 34); The deqi-feeling triggered in RA not in SA: Hervik et al. (53, 55), Crew et al. (59), Liljegren et al. (54), Oh et al. (60), Mao et al. (25, 34), Serra et al. (56).
Moreover, acupuncture is an intense interaction between patient and acupuncturist. Thus, differences in the procedure with less interaction in one arm may not lead to unblinding but may lead to a weaker response and thus explain a difference between RA and SA independent of a specific effect of acupuncture:
The deqi-feeling triggered in the RA group to find the exact puncture point: Hervik et al. (53, 55), Crew et al. (59), Liljegren et al. (54), Oh et al. (60), Mao et al. (25, 34), Jeong et al. (27), Serra et al. (56);
Open or hidden differences in setting and attention/attitude of the acupuncturist: Deng et al. (52), Hervik et al. (53, 55), Liljegren et al. (54), Crew et al. (59), Bokmand et al. (23), Oh et al. (60), Mao et al. (25, 34);
Individual points for patients or semi-standardized protocols: Crew et al. (59), Mao et al. (25, 34), Hershman et al. (24, 26), Serra et al. (56).
Figure 3 and Table S5 show these additional concerns. In fact, every study comparing RA and SA has at least one additional concern. Moreover, the positive effect in the RA arm may be described as an attention and interaction effect which is more than a placebo effect.
The inclusion of additional points of criticism may alter the assessment of the RoB tool for certain sub-items. Thus, if the additional criticisms of unblinding were included in six studies, the assessment of the domain Measurement of the outcome would change from low risk to high risk of bias (23, 26, 53, 55, 56, 58).
Some authors try to detect unblinding by asking participants after the intervention, whether they believe to have been in the RA or SA group. Yet, this question does not demask the influence of a more intense interaction as in both cases the patient may believe to be in the verum arm.
Waitlist control groups can also be subject to bias, as patients may expect their symptoms to improve only after this waiting period. In Mao et al. (25) and Hershman et al. (24), patients in the WLC group were able to receive 10 RA sessions after their follow-up, but there is no information on how many patients took advantage of this.
Comparison of methodological assessments from the SRs. While the SRs use different tools to assess methodological quality, this does not explain the large heterogeneity in the assessment of the same study rated by several SRs. Table S7 shows the methodological quality of the studies assessed, even ranges from low concerns to high concerns for individual items. There is no study for which the RoB ratings by two different SRs are exactly the same. While some SRs are more rigorous [e.g., Pan et al. (51)] and others are less rigorous [e.g., Ausanee et al. (39)], our assessment including our additional points of concern, led to high concerns about the overall risk of bias in 25 RCTs. Table S7 provides an overview of the RoB-assessments of all SRs.

Discussion
In our systematic assessment of published SRs and RCTs, we were able to show that the quality of publication from both types is rather low. All RCTs and SRs have to our mind severe methodological drawbacks which entail a high risk of bias and lower confidence in the results. Furthermore, the additional points introduced by us often lead to further downgrading of the evaluation of reliability. Comparing the RoB assessment of the studies in the different SRs, the assessment of the single studies differs between the reviews. Also, it is questionable whether the results of the methodologically poor RCTs are reliable when summarized in SRs.
There are 15 studies that address the effect of acupuncture on hot flashes, but the methodological quality of the studies is low: All studies have high concerns in the RoB assessment. Thus, despite this seemingly large number of studies, none can provide clear evidence whether acupuncture has an efficiency that is more than a placebo or attention/interaction effect. Some of the studies reported positive results of acupuncture on hot flashes. Considering musculoskeletal symptoms, fewer RCTs are available with all seven studies received a rating of high concerns, among those four which reported positive results for acupuncture. Accordingly, there is no reliable evidence on this issue.
Typical drawbacks of RCTs on acupuncture against menopausal symptoms in cancer patients are the usually small numbers of participants, a large percentage of drop-outs, missing outcome data, or impossibility of blinding in case of two different interventions (e.g., acupuncture and hormone therapy). Also, in some cases, the people who analyzed the data were not blinded and may have unknowingly influenced the results. In addition, in some studies, data were analyzed on a per-protocol basis and not as intent to treat.
Moreover, it is worth taking a closer look at some other items that address specific problems in acupuncture studies. Due to the nature of an acupuncture procedure, interactions between acupuncturist and patient are to be expected, as well as risks of unblinding during the study. On the one hand, blinding the acupuncturist is rarely possible, as a trained acupuncturist knows which points are real and which are false. This may also lead to some differences in the interaction of acupuncturist and patient which may lead to a stronger effect in case the acupuncturist is working with more conviction in case of RA than SA. In fact, there are several differences between RA and SA that may lead to unblinding or at least doubt on the part of the patient as to which arm the patient is in. A different depth of insertion of the points between RA and SA as well as different numbers of needles inserted or even largely different location may be detected by patients with some knowledge on acupuncture. Also using semi-standardized or individual protocols in the RA group and standardized protocols in the SA group may have these effects. Different depths of insertion or the use of non-penetrating sham needles, which may be attached with tape, can also lead to unblinding. Even if patients do not notice this explicitly, in all these cases, the context of the treatment, the interaction between acupuncturist and patient are different which may entail differences in the outcome.
To our knowledge, we are the first to draw the attention to another difference in the interaction between acupuncturists and patients. In fact, in RA often a so-called deqi feeling is triggered. To find the exact point, the patient has to tell the acupuncturist when the feeling occurs after the needle is placed. This interactive component actively involves the patient in the treatment and is very likely to evoke the idea of a particularly precise and individual point localization with the patient’s help. The patient himself contributes to the treatment. This raises the question of whether the insertion of needles at defined acupuncture points has an original mechanism of action or whether the effects may be explained solely by the application and interaction.
To assess this effect, studies with 3 arms comparing RA with SA and a control group without intervention are needed. In our SR, the five included studies with these conditions reach different results (23-26, 34). However, these results should be interpreted with caution due to the low methodological quality. In two studies, both EA and SA are better than the control groups, but there are no statistically significant differences between the EA and SA groups (25, 34), in one study the blinding is not certain (25). In three studies, the RA group is better than the SA group (23, 24, 26), with one showing participant unblinding (24) and in the other the SA group is never better than the no treatment group, although unblinding cannot be ruled out here either (23). The deqi feeling was asked from the patient in nine of the 22 included RCTs (25, 27, 34, 53-56, 59, 60). In seven RCTs, this assessment occurred only in the RA arm, likely resulting in greater attention for RA patients compared to those in the SA group (25, 34, 53, 55, 56, 59, 60). This introduces a bias favoring the RA group, potentially explaining why the RA group outperformed the SA group in four studies (53, 55, 56, 59). For the other three studies, no significant differences were found between the groups (25, 34, 60). In one study, RA with deqi stimulation was compared with no treatment, and in this study too, RA was significantly better than the comparison group (27). Only, in one of the studies, the triggering of a deqi sensation was simulated in the SA group (54). However, in the study the simulation differed significantly from the triggering of the deqi sensation in the RA arm (non insertive stimulation superficial) and the cooperation of the patient in the RA group was significantly more important than in the SA group (54).
If the additional points of criticism that have been identified are taken into account in the RoB tool, the rating of a domain would change in 6 studies. This shows that it can be important to identify the additional points of criticism and take them into account in the RoB assessment. RCTs of poor quality can thus be identified even more reliably. With regard to the included RCTs, it can be concluded that despite 22 included studies, no conclusions can actually be drawn about the effectiveness of acupuncture on menopausal symptoms.
In order to improve the methodological quality of studies on acupuncture, the STRICTA checklist was developed in 2001 and revised in 2010 (61). The details of acupuncture to be defined are listed in 6 points. This new standard was valid at the time the newer studies were planned and conducted, but only six of the 22 RCTS included refer to STRICTA in their methods section (24, 26, 28, 30, 59, 60) (see Table S8). Despite the focus on STRICTA, all six studies in the RoB have a rating of “high concerns”. The additional assessment points developed by us also lead to additional concerns in four of the six studies (24, 26, 59, 60). It is therefore questionable whether STRICTA has an impact on quality, especially since the main points are not addressed. Although the points and the insertion, including the triggering of the deqi sensation, should be described, STRICTA does not provide any information on actions to reduce the risk of bias in this regard. For example, the election of deqi in the RA group is described according to STRICTA, but no comparable procedure is used in the SA group in all four studies (24, 26, 59, 60).
To summarize, a specific adaptation of the RoB tool will be helpful to assess the reliability of acupuncture studies. An additional list, including the problems we outlined in our review, would be a practical addition to the RoB for studies on acupuncture. This most probably will also reduce the high heterogeneity in SRs on acupuncture assessing the same studies in the future and may contribute to a better evaluation of the evidence on an interactive intervention. Considering the effectiveness of acupuncture for menopausal symptoms due to the low methodological quality, no final conclusion can be drawn.
Firstly, regarding the acupoints the studies are highly heterogeneous with different points and different methods of application. Regardless of the methodological evaluation of the studies, it would also be almost impossible to find a clear concept for acupuncture in menopausal symptoms in breast cancer.
Second, the results reported are primarily self-reported and not objectively assessed. Thus, it is impossible to differentiate between an improvement due to a specific effect of acupuncture or due to the interaction between the patient and the acupuncturist. As this interaction is stronger in the verum as in sham acupuncture in the trials we assessed, this is not only a blinding problem as patients may well be kept blinded in the sense of not actively discerning in which group they are. But due to the interaction on the deqi-sensation, the interaction is quite more intense in the verum group. We raised this question in a former publication (62).
This leads to the question of how it would be possible to effectively balance a study. One established procedure is to have an inexperienced practitioner perform the acupuncture. Yet, the deqi-sensation, from the traditional Chinese medicine point of view, most probably may only be elicited at “true” acupoints; from a skeptical point of view, eliciting this sensation on other points of superficial innervation would be possible and would allow for the same intensity of interaction.
Study limitations. Limitations of this work are that only articles in German and English have been examined. It was also only possible to search databases in these two languages. Studies on acupuncture have been reported in other languages as well, mainly in Chinese language. We were not able to assess, these studies. Another limitation is that we did not include all symptoms that fall under the symptom complex of menopausal symptoms.

Conclusion
Two conclusions may be drawn from our review. First of all, so far, no RCT proves a specific effect of acupuncture on menopausal symptoms due to endocrine treatment in cancer patients. Indeed, the positive effect observed in the RA arm of some studies can be described less as a placebo effect and more as an attention and interaction effect. Consequently, the widespread use of acupuncture in cancer care and even positive recommendations in guidelines must be questioned. It might be argued that an improvement due to unspecific effects is worthwhile, yet acupuncture entails costs and time by the patient spent to visit the acupuncturist. Maybe more important is that there is an ethical question whether patients must be informed on the lack of evidence to make an informed decision.
Secondly, in complex treatments including interactions between therapist and patient, assessing the methodological quality may need additional points to consider which go beyond the RoB tool. While the STRICTA list is a first try to improve the quality of studies, add-ons regarding RoB might help authors of SRs and MAs to thoroughly assess the published studies.

Supplementary Material
https://github.com/kiaraluisakreher/Supplement

Conflicts of Interest
The Authors declare no conflicts of interest.

Authors’ Contributions
K.K. wrote the main manuscript text; the study was supervised by J.H. and J.D; J.B. and J.B. reviewed the text.

Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.