Harmonizing Formula Prescription Patterns for Chronic Kidney Disease: A Population-Based Cross-Sectional Study

Harmonizing formulas have been demonstrated to be associated with reduced risk of end-stage renal disease in patients with chronic kidney disease (CKD). However, the target population and indications of harmonizing formulas in CKD remain unknown. We conducted a population-based cross-sectional study to explore factors associated with harmonizing formulas prescription. Patients who had been prescribed harmonizing formulas after CKD diagnosis were dened as the using harmonizing formulas group. Disease diagnoses for harmonizing formula prescriptions and patient characteristics related to the prescriptions were collected.

Several experimental animal studies have concluded that certain herbs (e.g., Radix Salvia miltiorrhizae, Rhubarb) in CHM have potential anti-in ammatory and anti brotic effects and could be an agent in renal brosis therapy (5,6). However, these ndings are less applicable in clinical practice because illnesses in humans are often more complex than those in mice. For example, CKD in humans may have different etiologies and be accompanied by noncommunicable diseases. Therefore, observational human studies of CHM treatment effects provide an opportunity to understand current practices in CHM therapy and could help in exploring new therapeutic formulas for use in CKD. Studies and our previous data have demonstrated the potential bene cial effects of CHM on CKD outcomes (7)(8)(9)(10), highlighting that further understanding of the target population in CKD for treatment with speci c formulas is warranted.
Harmonizing formulas are commonly used in treating disorders caused by contrasting illnesses including shaoyang (half interior and half exterior) syndromes, liver and spleen disharmonies, and intestine and stomach disharmonies (11). The effectiveness of harmonizing formulas has been widely recognized in reducing depressive syndromes (12,13) and improving survival in patients with stroke and cancer (14).
However, applications for patients with CKD remain relatively less understood. Our study explored disease categories for which prescribing harmonizing formulas could be appropriate and independent factors associated with harmonizing formula prescriptions in patients with CKD.

Data Source and Subjects
The Taiwan government launched the National Health Insurance (NHI) program in 1995. It covers 99.6% of Taiwan's population and has service contracts with 93% of health care institutes. NHI reimburses medications (Western medicine and CHM), ambulatory, emergency, dental, and inpatient care after health care institutes complete medical services and uploads claims data. The claims data are further managed by the Taiwan National Health Research Institutes for inclusion in the National Health Insurance Research Database (NHIRD) and are available for academic research.
We conducted a population-based cross-sectional study by using the Longitudinal Health Insurance Database 2000 (LHID2000), a subset of the NHIRD. The LHID2000 contains the data of 1 million randomly sampled patients who were NHI bene ciaries in 2000. The randomly sampled patients exhibit similar distributions in age, sex, birth year, and average insured payroll-related amount with the general population.

CKD
The detailed methods of this study were in our previous study (10). In brief, we identi ed patients who received new diagnoses of CKD between 2000 and 2005 by using the frequency of appearance of speci c International Classi cation of Diseases, Ninth Revision, Clinical Modi cation (ICD-9-CM) codes within one year (15). Although laboratory data were lacking in these databases, we can reasonably assume that most of these patients had stage 3-5 CKD [estimated glomerular ltration rate (eGFR) < 60 ml/min/1.73 m 2 ] on the basis of regional hospital data using the same algorithm. We excluded patients who were aged < 18 or ≥ 85 years (n = 2, 673), had cancer (n = 1,680) or underwent dialysis before receiving a CKD diagnosis (n = 28), or had received any CHM prescription within 1 year before diagnosis (n = 11,104). Therefore, the cohort was suitable to assess for determining factors associated with CHM prescription.

Harmonizing Formulas
All CHM prescriptions from CKD diagnosis to the start of dialysis or the end of 2008 were collected. We distinguished these CHM prescriptions into harmonizing formulas and other formulas according to approaches suggested in textbooks, with minor modi cations (11,16). To detect potential indicators for using harmonizing formulas, disease categories for prescribing harmonizing and other formulas were compared on the basis of a patient's rst ICD-9-CM diagnosis code of the prescription and classi ed into various disease system and health problem groups. These systematic diseases and related health problems can be divided into infections and parasitic diseases (ICD-9-CM: 001-139); neoplasms (140-239); endocrine, nutritional, and metabolic diseases, and immunity disorders (240-279); diseases of the blood and blood-forming organs (280-289); mental disorders (290-319); diseases of the nervous system and sensory organs (320-389); diseases of the circulatory system (390-459); diseases of the respiratory system (460-519); diseases of the digestive system (520-579); diseases of the genitourinary system

2-1. Statistical Analysis
Continuous and categorical data were expressed as mean ± standard deviation or median (interquartile range) and percentage, respectively. Signi cant differences in patient characteristics between the harmonizing formula use and nonuse groups were evaluated using an independent t test for continuous variables and χ 2 test for categorical variables. We used the proportion of disease categories to determine the differences in disease treatment between using harmonizing formulas and other formulas. In addition, a multivariable logistic regression with forced entry of all collected factors was performed to identify the independent factors of harmonizing formula prescription. Results of the logistic regression were represented as odds ratios (ORs) and 95% con dence intervals (CIs). All statistical operations were performed in SAS (version 9.4, SAS Institute, Cary, NC, USA). A P value < 0.05 was considered signi cant.

Patient Characteristics by Use of Harmonizing Formulas
We included 24,971 patients who received new diagnoses of CKD, and 21% of these patients were prescribed harmonizing formulas. Compared with patients in the harmonizing formulas nonuse group, patients in the group using harmonizing formulas were signi cantly more likely to be young, female, and living in central Taiwan and urban areas. They were signi cantly more likely to have high insurance amounts, fewer comorbidities (acute coronary syndrome, diabetes, hypertension, hyperlipidemia, COPD, or cerebrovascular disease), less disease severity, and less use of diabetic and antihypertensive drugs, but more use of NSAIDs, analgesic drugs other than NSAIDs, and antilipid drugs.

Disease Categories for Prescribing Harmonizing Formulas
The disease categories for prescribing harmonizing formulas and the other types of Chinese herbal formulas in patients with CKD were compared and are listed in Table 2. The three most frequent disease categories for prescribing harmonizing formula in CKD were symptoms, signs, and ill-de ned conditions (24.5%); diseases of the digestive system (20.67%); and diseases of the musculoskeletal system (12.9%).
Similar frequencies of disease categories were observed for other formula prescriptions in patients with CKD. Notably, an increased proportion of patients with diseases of the digestive system (6.39% difference) and a reduced proportion of those with diseases of the respiratory system (− 5.5% difference) received harmonizing formula prescriptions compared with other formula prescriptions in this analysis. The differences in characteristics among groups were compared using χ 2 tests for categorical variables and independent t tests for continuous variables. A P value of < 0.05 was considered statistically signi cant.

Prescription Frequency of Constituent Herbs in Harmonizing Formulas
As Fig. 1 presents, the three most frequently prescribed herbs for harmonizing formulas during the observed period were Jia Wei Xiao Yao San, Shao Yao Gan Cao Tang, and Xiao Chai Hu Tang. The most frequent disease categories for Jia Wei Xiao Yao San prescriptions were symptoms, signs, and ill-de ned conditions (n = 2,143 times); genitourinary disease (n = 1,767 times); and diseases of the digestive system (n = 1,157 times). The most frequent disease category for Shao Yao Gan Cao Tang prescriptions was diseases of the musculoskeletal system (n = 1,876 times). Furthermore, the most frequent disease category for Xiao Chai Hu Tang prescription was diseases of the digestive system (n = 1,056 times).

Discussion
The current study demonstrated that one-fth of patients with CKD have been prescribed harmonizing formulas, which were associated with risk reduction of ESRD in our previous study (10). Symptoms, signs, and ill-de ned conditions; diseases of the digestive system; and diseases of the musculoskeletal system were the three most frequent disease classi cations for prescribing harmonizing formulas.
Patients with CKD who were young, female, had high premiums, lived in central or southern Taiwan or urban areas, did not have comorbidities (acute coronary syndrome, diabetes, hypertension, or cerebrovascular disease), had lower disease severity, and used NSAIDs and analgesic drugs other than NSAIDs were more likely to have harmonizing formulas prescriptions.
Nearly half of incident CKD patients used conventional CHM. Harmonizing formulas accounted for 46.1% of prescriptions. Although the e cacy of harmonizing formulas in reducing depression and improving survival in patients with liver cancers and systemic lupus erythematosus have been reported in studies (14,17), few studies have paid attention to prescription patterns and outcomes in treating patients with CKD (18,19). , and our previous ndings, the current study indicated that one harmonizing formula, Jia Wei Xiao Yao San, is the main prescription for patients with late CKD. This formula potentially improves patient renal function after short-or long-term observation (10,19). Jia Wei Xiao Yao San is traditionally used to relieve stagnation in liver qi, reduce depression symptoms, and improve spleen qi de ciency. Although the mechanisms of delaying ESRD are complex and worthy of further study, Jia Wei Xiao Yao San may affect renal clinical outcome by improving depressive symptoms, which are a novel predictor of accelerated eGFR decrease, dialysis therapy initiation, death, or hospitalization (10,20). We are unsure of the effect harmonizing formulas have in relieving uremic symptoms or slowing renal progression.
We attempted to use the ICD-9-CM to as a reference for the indications of harmonizing formula prescription in CKD patients. CHM has its own historical and systematic philosophy of symptom differentiation (Bian Zheng) to assess, explicate, diagnose, and treat patient symptoms. Numerous symptoms in CKD patients not requiring dialysis are either unrecognized or suboptimally managed by clinical care workers (21,22). assessed 283 patients with stage 1-5 CKD and reported that tiredness (81%; 95% CI: 76.0-85.6), sleep disturbance (70%; 64.3-75.3), and pain in bones or joints (69%; 63.4-74.6) were the most common symptoms regardless of CKD stage. Loss of appetite, nausea, vomiting, fatigue, and edema are common symptoms in late-stage CKD, which are similar to "spleen de ciency syndrome" in CHM and tend to involve the digestive system. In addition, local pain, weakness in the loin and knee, and calf cramps are frequently encountered CKD symptoms, which are similar to "liver-kidney insu ciency syndromes" in CHM and tend to involve the musculoskeletal system. Shao Yao Gan Cao Tang, a frequently prescribed harmonizing formula in CKD, is used to relieve muscle pain or skeletal muscle tremors in Japan and China (23). However, additional bene cial effects of Shao Yao Gan Cao Tang on renal health remain unknown and warrant further investigation.
Female sex, low prevalence of comorbidity, and high use of analgesic drugs were associated with higher prescription frequency for harmonizing formulas in CKD. Although the causal relationships are di cult to establish in this study, a possible explanation is that prevalence of pain for females is high, and they are more likely to be aware of pain and receive relevant treatments (24,25). NSAIDs are commonly used for pain control in clinical practice, but caution should be exercised when they are applied in CKD because they can induce more severe renal injuries (26). How harmonizing formulas interact with NSAIDs for pain control in patients with CKD remains unclear. Thus, more research is required on this combination therapy to study its e cacy in pain control in CKD and preventing further renal injury.
Our study has some advantages. First, assessments of Chinese herbal formulas in the study were drawn from a nationwide health insurance database with highly comprehensive records of CHM prescriptions. Second, CHM in this study was prescribed by quality assurance physicians who were educated in the same system and accredited by Taiwan's government; this strengthens the reliability of symptomatic differentiation and accuracy of disease diagnostic coding. Third, Taiwan's NHI is one of the few national insurance programs that reimburses both Western medicine and CHM, providing an opportunity to explore the foundational philosophies of these two different modes for treating certain diseases. However, some limitations must be declared. First, the lack of laboratory and patient-reported data in the claims database prevents us from exploring the possible mechanisms of harmonizing formulas on health outcomes such as emotion, pain, and renal function. In addition, using the ICD-9-CM diagnosis system to identify disease classi cation of CHM prescriptions may not have accurately re ected the indications of CHM formulas. However, we believe that using the main categories of diseases partially represented the CHM diagnostic system. Finally, our results were from the NHI program having high accessibility, which may limit their generalizability.

Conclusions
In conclusion, this study determined that harmonizing formulas are commonly used in treating CKD for symptoms, signs, and ill-de ned conditions; genitourinary diseases; and digestive system diseases. Patients who were young, female, had fewer comorbidities, and used analgesic drugs were more likely to be prescribed harmonizing formulas, which suggests that more research on the e cacy of Western medicine and CHM in patients with CKD and these characteristics is required.

Consent for publication
The Institutional Review Board of Kaohsiung Medical University Hospital exempted from the requirement of informed consent because the personal privacy information within the analyzed data have been completely de-identi ed and allowed for improvement of public interest.

Availability of data and materials
All data generated or analysed during this study are included in this published article.

Con icts of interest
None of the authors have any con icts of interest to declare.  Prescription frequency of constituent herbs of harmonizing formulas categorized by systematic disease and related health problems in patients with chronic kidney disease.