Prevalence and patterns of multi-morbidity in the productive age group of 30-69 years: A cross-sectional study in Pathanamthitta District, Kerala.

Background: Multi-morbidity is the coexistence of multiple chronic conditions in individuals. With advancing epidemiological and demographic transitions, the burden of multi-morbidity is expected to increase India. Methods: A cross-sectional representative survey was conducted among 410 participants aged 30-69 years in Pathanamthitta District, Kerala to assess the prevalence of multi-morbidity. A multi-stage cluster sampling method was employed to identify households for the survey. We interviewed all eligible participants in the selected households. A structured interview schedule was used to assess socio-demographic variables, behavioral risk factors and prevailing clinical conditions. We used the PHQ-9 questionnaire for depression screening. Further, we conducted active measurements of both blood sugar and blood pressure. Multiple logistic regression was used to identify variables associated with multi-morbidity. Results: Overall, the prevalence of multi-morbidity was 45.4% (95% CI: 40.5-50.3%). Nearly a quarter of the study participants (25.4%) reported only one chronic condition (21.3-29.9%). Further, 30.7% (26.3-35.5), 10.7% (7.9-14.2), 3.7% (2.1-6.0) and 0.2% reported two, three, four and five chronic conditions, respectively. At least one person with multi-morbidity was present in around seven out of ten households (72%, 95% CI: 65-78%). Further, one in five households (22%, 95% CI: 16.7-28.9%) reported more than one person with multi-morbidity. Diabetes and hypertension was the most frequent dyad (30.9%, 95% CI: 26.5-35.7%), followed by hypertension and depression (7.8%, 95% CI: 5.5-10.9%). Diabetes, hypertension and ischemic heart disease was the common triad in males (8.5%, 95% CI: 4.8-14.1%), while it was diabetes, hypertension and depression (6.9%, 95% CI: 4.2-11.1%) in females. Age, sex, and employment status were associated with multi-morbidity. Conclusion: Multi-morbidity is prevalent in one of two participants in the productive age group of 30-69 years. Further, seven of ten households have at least one person with multi-morbidity. The high burden of multi-morbidity calls for integrated management strategies for multiple chronic conditions.


Introduction
Globally, several countries are going through an epidemiological transition in which degenerative diseases are displacing infections and nutritional disorders as the major cause of mortality and morbidity 1 . An increasing trend of life expectancy due to good living conditions, increased income per capita, advances in health care and education are reported globally 2,3 . However, the added years of life due to increase in life expectancy are often complicated by poor health and disability, resulting from the high burden of chronic diseases, especially non-communicable diseases (NCDs) 4 . Available evidence suggests that the majority of patients diagnosed with chronic diseases have more than one chronic co-existing condition 5 . Co-existence of more than one chronic condition in the same individual is known as multi-morbidity 6 .
Prevalence estimates of multi-morbidity vary widely depending on method of data collection 4 , age of the population studied, definition of chronic condition, definition of multi-morbidity, and number and types of chronic conditions included in the analysis. In addition to the complexities related to the definition, there is a lack of an explicit validated tool to measure the components of multi-morbidity uniformly for reporting the magnitude of diseases 7 . However, available evidence indicates that prevalence, pattern and complexity associated with multiple chronic conditions have been increasing over the years 8 .
The increasing burden of multi-morbidity may cause considerable financial burden, especially in low-and middle-income countries (LMICs). If preventive and management strategies are not integrated and coordinated across multiple chronic conditions, it can lead to duplication and health system inefficiency. Data from LMICs on multi-morbidity are very limited. Further, India does not have population level estimates on multi-morbidity. Given the rising burden of NCDs in India, with the onset of disease at least a decade earlier than their Western counterparts 9 , it is important to characterize the multi-morbidity pattern in this population.
The state of Kerala in India is in an advanced stage of epidemiological transition as compared to other states in India 10,11 . Therefore, it is highly likely that the prevalence of multi-morbidity will be higher in Kerala. However, very limited data on prevalence of multi-morbidity are available from Kerala. A better understanding of the epidemiology of multi-morbidity is crucial for re-organization of health care services to provide integrated care for multiple chronic conditions. The objective of our study was to assess the community level prevalence and pattern of multi-morbidity in Kerala.

Ethical statement
The Institutional Ethics Committee (IEC) of Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum approved the study (SCT/IEC/1455/NOV-2019). The questionnaire interview and all measurements were conducted after obtaining a written informed consent from each study participant. Privacy was ensured during the time of interview and confidentiality of all the information collected was maintained. The participants agreed to report or publish the data collected during the study except any information that could lead to the identification of any individual, by signing a consent form prepared in local language. The participants had the freedom to refuse participation at the beginning or during any stage of data collection.

Study design
A cross-sectional survey was conducted in Pathanamthitta district of Kerala, India using a structured interview schedule.

Study setting
Pathanamthitta is a southern district of Kerala with the highest proportion of elderly population (18%), lowest total fertility rate (1.3) 12 and highest literacy rate (96.9%) 13 . The population size of Pathanamthitta district is close to 1.2 million as per the 2011 census 14 . A high prevalence of diabetes is also noted in Pathanamthitta district 15 .

Study population
We included eligible participants aged 30 to 69 years. The participants were residents of rural areas of Pathanamthitta district for a minimum period of one year. Those who did not give informed consent, were physically or mentally not in a condition to answer the questions and/or undergo clinical measurements as part of the study, or were pregnant or lactating within six weeks post-partum were excluded. All eligible participants in a selected household who satisfied the inclusion criteria were included in the study.
There are eight community development blocks (CDB) in Pathanamthitta district. The CDBs were the primary sampling units. Of these eight CDBs, three were randomly selected by simple random sampling. We used computer generated random numbers for the random selection of CDBs. In the second stage, four panchayats were selected from each CDB using computer generated random numbers (a total 12 of Panchayats were selected from three blocks). From these 12 panchayats, one ward each was randomly selected using computer generated random numbers. Further, 16 to 18 alternate houses were visited in each ward. One of the authors (RC) was responsible for the selection of households and participants for the study. After locating the center and the main junction of the ward, by pen rotation, the first household was identified. Subsequently, every alternate household was visited. In case the selected house was empty or locked, the next nearest one was visited. All the houses visited were on the right hand side of the

Amendments from Version 1
We have made amendments to the data on ' employment status' presented in Table 1. We have highlighted that in the multivariate regression model, a sandwich variance estimator was used for accounting for clustering. We have also added the definition of marital status in the footnotes of Table 3.
Any further responses from the reviewers can be found at the end of the article REVISED road. All the eligible participants in the selected households were interviewed. In case any of the potential participants were not present in the house during the time of interview, one more attempt was made at a later time point.

Data collection
A medically qualified primary care doctor (RC) conducted the data collection from all participants. A community health worker (ASHA) assisted RC in the survey. A structured interview schedule 16 was used to assess variables such as sociodemographic factors, behavioral risk factors and prevailing clinical conditions. We used the translated patient health questionnaire (PHQ-9) in Malayalam (local language in the state of Kerala) for screening of depression. Random capillary blood sugar (Onetouch Verio Flex Meter) and blood pressure (Omron Blood pressure monitor-upper arm BP7100) measurements were also undertaken. Three readings were recorded at five minutes intervals in the non-dominant arm by getting the participant to sit comfortably for five minutes and keeping the machine at the same level of his/her heart. The average of the three values was calculated and considered as the blood pressure of the participant. Elderly (above 60 years) participants were asked about their living arrangements.
Females were asked about their history of menopause. The period of data collection was from 01.01.2020 to 28.02.2020.

Definitions
The operational definition of 'multi-morbidity' was 'co-existence of more than one of the 11 listed chronic conditions in the same individual'. Hypertension was defined as per Joint National Committee-7 guidelines 17 . Depression was defined as a PHQ-9 score of 10 or above. Diabetes mellitus was defined as a random capillary blood glucose value above 140mg/dl. The remaining conditions viz ischemic heart disease, heart failure, stroke, chronic kidney disease, chronic obstructive airway disease, arthritis, thyroid hormone disorders and cancer were self-reported by the participants and verified with corresponding patient-held medical records, which were present with them. The patient-held medical records included consultation details, diagnostic details and treatment details. Those who ever used any form of tobacco were defined as tobacco users and those who ever used alcohol were defined as alcohol users. Recommended level of physical activity was defined as walking/jogging/engaging in sport for at least 30 minutes a day on five days in a week. A household was defined as a group of persons who live together and share a common kitchen. The age of starting formal education in the state of Kerala is six years. Low education group were those who studied up to 7 th standard. Those who studied above 7 th standard were defined as high education group.
Sample size A prevalence of 30% was anticipated 18 and the precision assumed was 6%. A design effect of 1.5 was assumed, as we employed a multi-stage cluster sampling to identify the participants. We accounted for 10% attrition. The sample size required was estimated as 403 and it was rounded off to 410.
Data management and data analysis Data were collected using paper forms and later entered into Microsoft Excel 2013 (v 15.0) sheets 16 . Data cleaning and data analysis were done using R software version 3.6.3 19 . We used the R packages "summarytools' and 'ggplot2'. Continuous variables were presented as mean and standard deviation and categorical variables as proportions. The overall prevalence and prevalence across socio-demographic variables were calculated with 95% confidence intervals (CIs). Logistic regression analyses were performed to estimate odds ratios (ORs) of various socio-demographic factors associated with multi-morbidity. We present both un-adjusted and adjusted multi-variate model (with sandwich variance estimator) results.

General characteristics
Out of the 430 participants approached, we collected data from 410 participants (95% response rate) 16 . The mean age of the study population was 53 (SD=11.7) years ( Table 1). Three of five participants in the study population were females (59.8%). Females were younger (mean=52, SD=12.2 years) compared to males (mean=55, SD=10.7 years). Nearly one-third of the participants (32%) reported education up to seventh standard. Nearly half (49%) of the study population were homemakers (49%). One of five participants (20%) were ever users of tobacco. Alcohol use (ever use) was prevalent in 14.6% of the study population. Recommended level of physical activity was reported in one of five participants (20%).
Around seven in ten households (72%) reported at least one person in the household with multi-morbidity. Additionally, one in five households (22%) reported more than one person with multi-morbidity ( Figure 1B). Overall, 28% of the households did not report any one with multi-morbidity.

Multi-morbidity in sub-groups of population
The prevalence of multi-morbidity increased from 3.2% in the 30-39 age group to 69.7% in the 60-69 age group (p<0.001).
More than half of the study participants (53.1%) in the 50-59 age group reported multi-morbidity (Table 3). In the multiple logistic regression analysis, a 10% increase in   Among the 245 female participants, 63.3% had reached menopause. Among those who had reached menopause, the prevalence of multi-morbidity was 66.5%. However, the prevalence of multi-morbidity was 16.7% among females who did had not reached menopause.

Discussion
Highlights Multi-morbidity, the co-existence of multiple chronic conditions in the same individual, is prevalent in one of two adult individuals in the age group of 30-69 years in Pathanamthitta district of Kerala. The prevalence of multi-morbidity increases with advancing age. We also document that seven in ten households have at least one individual with multi-morbidity. Further, one in five households report more than one person with multi-morbidity. Diabetes and hypertension are the most common co-existing conditions. However, we find differences in the second and third most common dyads and triads of co-existing chronic conditions between males and females. Multi-morbidity assessment in management of chronic conditions is essential given its high prevalence and the differences in health care utilization and quality of life in individuals with different combinations of co-existing chronic conditions 20 .
Comparison of multi-morbidity prevalence with other studies Direct comparison of prevalence of multi-morbidity across studies is almost impossible due to varying definitions of multi-morbidity, methods of data collection and the number of conditions included. In a systematic review, the prevalence of multi-morbidity ranged from 4.5% to 83% in the South Asian population 21 . Additionally, in a previous study conducted in Kerala among the age group of 45 years and above, the prevalence of multi-morbidity was substantially lower than our study estimates 22 . Most of the studies on multi-morbidity in India and other LMICs rely on self-reported data on chronic conditions from the participants and are often subjected to several biases. In our study, we incorporated active measurements for diabetes, hypertension and depression assessment. Further, the medical records of the participants were verified by a medical doctor to confirm the self-reported data on other chronic conditions. The number of chronic conditions included in our assessment for multi-morbidity was also higher than other studies conducted in LMICs.

Multi-morbidity assessment at the household level
The household is an important unit of intervention for chronic disease prevention and control. Although the data on the *Only variables that showed statistical significance (p<0.05) in unadjusted analyses were entered into adjusted multi-variate analysis. **Analysis was done only for male participants as females did not report alcohol or tobacco use. MM, multimorbidity; OR, odds ratio. ***Currently not married means either unmarried or divorced/separated or widowed prevalence of multi-morbidity at the household level are limited, available data suggest strong concordance of chronic conditions within the co-residing adults 23 . Existence of at least one person with a chronic disease among seven out of ten households in Kerala is alarming. Further, multi-morbidity was prevalent in more than one individual in one of five households. The alarming prevalence at the household level calls for innovative models for prevention and control of multi-morbidity.
Family-based models such as the Programme of Lifestyle Intervention in Families (PROLIFIC study) 24 may be more appropriate to reduce the burden and progression of multi-morbidity in LMICs. A family centered approach for lifestyle changes and self-care for cardiovascular risk reduction, and involvement of non-physician health workers for care-coordination were the key strategies in the PROLIFIC study. It is envisaged that in a family focused approach, the proposed lifestyle changes and self-care strategies are more achievable and sustainable for both the individuals and their family members 25 .

Multi-morbidity prevalence in sub-groups and their implications
The prevalence of multi-morbidity increases with age. In our study, the highest prevalence of multi-morbidity was among the age group of 60-69 years. Similar findings were reported in studies from other countries 21,26-28 . Although seven out of ten participants in the 60-69 age group reported multiple chronic conditions in our study, it was evident that multi-morbidity is a substantial problem even in the younger population. The relatively high burden of multi-morbidity in the most productive age group is a cause for concern as it has serious implications in terms of productivity loss, higher health care utilization 20 and health expenditure 29 .
In our study the prevalence of multi-morbidity among females was higher in comparison to males. Several other studies from India and other Asian countries reported a higher prevalence of multi-morbidity in females 22,29,30 . Further, high prevalence of multi-morbidity among unmarried, divorced/separated/widowed individuals are also reported in other studies 31,32 . Gender differences in multi-morbidity pattern and care utilization should be explored in detail in future studies.
There is contrasting evidence regarding the association of education or socio-economic status with multi-morbidity. Data from high-income countries suggest an inverse relationship with education 33 , while the relationship is positive and linear (prevalence increases with increase in educational level) in studies from LMICs 31 . Further, some studies from India and Kerala failed to establish any relationship between educational status and multi-morbidity 34,35 . Our study indicates a higher prevalence in individuals with relatively low education levels. However, on adjustment for other socio-demographic characteristics, individuals with less than primary school education had lower propensity for multi-morbidity as compared to individuals with more than primary school level education. Under-reporting of self-reported chronic conditions, especially in the low education group 36 , often results in positive associations of education with multi-morbidity. We recommend further investigation to profile the pattern of multi-morbidity according to attained educational status and related inequality.

Multi-morbidity pattern and their implications
The most common pair of coexisting chronic conditions in our study was diabetes and hypertension in males and females. This finding is consistent with the results of the study conducted by Singh and colleagues 27 . The second most common pair was hypertension and depression for females, while it was hypertension and ischemic heart disease for males. Similarly, the most common triad in females was diabetes, hypertension and depression, while in males it was diabetes, hypertension and ischemic heart disease. Higher health care utilization and poor quality of life 20 in individuals with multiple chronic conditions calls for integrated prevention and control strategies to address the burden of multi-morbidity.

Strengths and limitations
In our survey, we achieved a high response rate (95%) and incorporated measures for active screening of under-diagnosed chronic conditions such as hypertension, diabetes and depression. Further, a medically qualified primary care physician reviewed the medical records of all participants to confirm the self-reported status of other chronic conditions. However multi-morbidity assessment was done by simple counting and the severity of the conditions were not taken in to consideration. Random blood sugar for identification of individuals with diabetes may have under-estimated the true burden of diabetes. Association observed in our study does not imply causality due to the cross-sectional study design.

Conclusion
Multi-morbidity is a major public health problem in Kerala that affects almost half of adult individuals in the productive age group and seven out of ten households. Diabetes, hypertension, depression and ischemic heart disease are the most common co-existing conditions. However, the pattern of multi-morbidity is different across gender and other socio-economic groups. Future research is recommended to identify the progression of single chronic conditions to multi-morbidity over the life-course using prospective study designs.

Open Peer Review
Abstract: The abstract is concise, specific, representative of the article, and in the correct form, which accurately summarizes the essential information.
The first sentence in the abstract, "Multi-morbidity is the coexistence of multiple chronic conditions in individuals and families," is slightly different from the definition of multimorbidity given in the methodology. Better to avoid "families" from abstract.

○
The objective is not given in the abstract.

Methods:
The methodology of the study is thorough and presented clearly. The researchers have done a good job describing the study population, data collection, definitions of terms used, and sample size.
Since all the eligible participants in the selected households were interviewed, there is a possibility for the clustering of multiple respondents within the houses, which is not addressed in the analysis.

○
The details of the number of households surveyed were not given in the abstract. This is important in the context of the discussion on multi-morbidity assessment at the household level.

○
Only elderly participants were asked about their living arrangements. This needs further explanation.
○ Under the sub-title 'Data Collection,' "Malayalam" can be put in italics and bracket (local language), helping international readers.

○
The justification for selecting the 11 listed chronic conditions in the study were not given.

○
There is no mention of the type of treatment for the selected diseases.
○ Patients with multi-morbidity are at increased risk for polypharmacy. Since the authors verified the medical records, it would have been easy for them to get details of medication and the treatment from different systems.

Results:
The results are clearly described. Some points for consideration are as follows: The study used multi-stage cluster sampling. Correction for clustering was not performed in data analysis.
○ Depression was measured using PHQ-9. However, no detail on the prevalence of depression is presented in the manuscript.

○
The results derived from the logistic regression analysis presented in Table 3, age standardization would have helped estimate the true prevalence of multi-morbidity in this population, which would also help the readers to compare results from other studies.

○
The classification of employment status presented in Table 1 & Table 3 is different. I'd suggest making these consistent.

○
In Table 2, the authors gave details of common pairs of chronic conditions. However, I couldn't find the prevalence of individual diseases anywhere in the manuscript.

○
In Table 3, in employment classification, better to club the 'salaried' with the first category is 'daily wage/self-employed' so that it becomes two categories as employed and not employed.

○
In Table 3, marital status classification needs to be verified, specifically the term 'currently not married.' ○ Prevalence of the recommended level of physical activity was given in the manuscript. It would be better to analyze the recommended level of physical activity with multi-morbidity considering the association of physical activity with decreased risk of chronic conditions and mortality.

Discussion:
The discussion section is very well presented as it covers all of the components associated with the study. The conclusion is based on the study's findings, and the recommendations made would help future related studies.
In the results section, the association of menopause and multi-morbidity was given, and there is a significantly increased prevalence of multi-morbidity among those who reached menopause. However, no further discussion is seen in this association. Higher age of menopause women may be considered in this case.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Author Response 09 Dec 2020 Panniyammakal Jeemon, Sree Chitra Tirunal Institute of Medical Sciences and Technology, Trivandrum, India Thank you very much for the positive comments. We have included our detailed response to the comments in the section below:

Comment 1: The first sentence in the abstract, "Multi-morbidity is the coexistence of multiple chronic conditions in individuals and families," is slightly different from the definition of multimorbidity given in the methodology. Better to avoid "families" from abstract. ○ ○ The objective is not given in the abstract.
As per the suggestion we have remove the word "families" from the abstract. We have also added the main objective of the study in the abstract.

Comment 2a: Since all the eligible participants in the selected households were interviewed, there is a possibility for the clustering of multiple respondents within the houses, which is not addressed in the analysis.
We had included all eligible participants to assess the multi-morbidity burden at the household level. The household level multi-morbidity is a new concept and has important implications in prevention and control of chronic conditions. We had included 193 households (clusters) in the study and used a sandwich variance estimator in the regression model.

Comment 2c: Only elderly participants were asked about their living arrangements. This needs further explanation.
We were interested in generating some pilot data on the relationship between living arrangements and multi-morbidity burden in the elderly population. It was not part of the main objective of the study.

Comment 2d: Under the sub-title 'Data Collection,' "Malayalam" can be put in italics and bracket (local language), helping international readers.
We have inserted a bracket and mentioned that it is the local language in Kerala. Comment 2e: The justification for selecting the 11 listed chronic conditions in the study were not given. This is purely because of pragmatic reasons and based on consensus opinion from experts in the field.

Comment 2f: There is no mention of the type of treatment for the selected diseases. Patients with multi-morbidity are at increased risk for polypharmacy. Since the authors verified the medical records, it would have been easy for them to get details of medication and the treatment from different systems.
Yes. As the reviewer correctly pointed out the patients with multi-morbidity are at increased risk of poly-pharmacy. Although we collected details of medications, we could not analyze the data as there was no uniformity regarding the type of treatment especially between various systems of medicine. We will need to discuss this with experts in alternative systems of medicine to complete the data analysis.

Comment 3a: The study used multi-stage cluster sampling. Correction for clustering was not performed in data analysis.
We had included 193 households (clusters) in the study and used a sandwich variance estimator in the regression model. Comment 3b: Depression was measured using PHQ-9. However, no detail on the prevalence of depression is presented in the manuscript Prevalence of individual conditions are also presented in Figure 2. Comment 3c: The results derived from the logistic regression analysis presented in Table 3, age standardization would have helped estimate the true prevalence of multi-morbidity in this population, which would also help the readers to compare results from other studies. We agree that age-standardization would have helped for comparison of data with other studies. However, since it was a study conducted in one of the districts in Kerala, it may not be useful even if we age standardize to the world standard or Indian population. Comment 3e: In Table 2, the authors gave details of common pairs of chronic conditions. However, I couldn't find the prevalence of individual diseases anywhere in the manuscript.
Prevalence of individual conditions are also presented in Figure 2. Comment 3f: In Table 3

, in employment classification, better to club the 'salaried' with the first category is 'daily wage/self-employed' so that it becomes two categories as employed and not employed.
Thank you for this suggestion. We have made the changes as per the suggestion. Comment 3g: In Table 3, marital status classification needs to be verified, specifically the term 'currently not married. ' We have added a footnote to explain the definition of this variable.

Comment 3h: Prevalence of the recommended level of physical activity was given in the manuscript. It would be better to analyze the recommended level of physical activity with multi-morbidity considering the association of physical activity with decreased risk of chronic conditions and mortality.
The limitation of the study design made the assessment difficult as many of the participants started doing the recommended level of physical activity only after they have been diagnosed with one or more of the chronic conditions.

Comment 4: In the results section, the association of menopause and multi-morbidity was given, and there is a significantly increased prevalence of multi-morbidity among those who reached menopause. However, no further discussion is seen in this association. Higher age of menopause women may be considered in this case.
As the reviewer pointed out, there was a higher prevalence of multi-morbidity among females who attained menopause. However, after adjusting for age this difference was marginal.

Pallab K Maulik
George Institute for Global Health, New Delhi, Delhi, India It is good paper ascertaining multimorbidities in the south Indian state of Kerala (one district).
Overall, I have a few observations which are now not modifiable, but something that needs to be addressed in the discussion: The district chosen seems one of the more extreme districts of Kerala in many ways as already indicated by the author and Kerala itself is one of the odd states in India -health indicator wise. This does bring into question issues about generatlizability of the data and more discussion needs to be given to that. 1.
The study excluded pregnant and lactating women. Not sure about the rationality as it definitely takes out the possibility of perinatal depression and some other pregnancy related health issues. 2.
There is no information about the temporality of the occurrence of the diseases. If that data was not collected, one needs to include that in the limitations and comment on it. 3.
While the study may not be powered to provide gender-based analysis, some preliminary indications of trends could be useful, especially given that the results were affected by gender.

4.
The author may want to look into the Academy of Medical Science, UK report on multimorbidity.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Partly

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes We thank the reviewer for their positive comments. Our detailed response to the comments are given below;

Comment 1: The district chosen seems one of the more extreme districts of Kerala in many ways as already indicated by the author and Kerala itself is one of the odd states in Indiahealth indicator wise. This does bring into question issues about generalizability of the data and more discussion needs to be given to that.
We agree that the results may not be generalisable to other population with different population structure. We have therefore given the name of the district and the state in the title itself. Further, we have described the socio-demographic pattern of the selected district in the study settings.

Comment 2: The study excluded pregnant and lactating women. Not sure about the rationality as it definitely takes out the possibility of perinatal depression and some other pregnancy related health issues.
We agree that exclusion of pregnant and lactating women would have underestimated the prevalence of multi-morbidity. However, some of the chronic conditions like depression may be relatively transient (not very chronic) in this group of participants. Further, it is a standard methodology to exclude pregnant women from similar surveys.

Comment 3: There is no information about the temporality of the occurrence of the diseases. If that data was not collected, one needs to include that in the limitations and comment on it.
Yes. As the design of the study is cross-sectional, temporality cannot be ensured. We accept it as a limitation of this study design.

Comment 4: While the study may not be powered to provide gender-based analysis, some preliminary indications of trends could be useful, especially given that the results were affected by gender.
We have presented the common pairs of chronic conditions stratified by gender in Table 2.

Comment 5: The author may want to look into the Academy of Medical Science, UK report on multi-morbidity.
Thank you for this suggestion. We were motivated to conduct this assessment after reading the Academy of Medical Science, UK report on multi-morbidity.
Once again we thank the reviewer for his positive comments.