Gender prevalence of cardiovascular diseases in the geriatric population of India: A meta-analysis using R

Abstract

BACKGROUND

Population ageing is an important challenge for developed as well as developing countries due to the downward trends in mortality rates. The elderly population is increasing worldwide. Cardiovascular diseases (CVDs) are one of the most common diseases in the geriatric population. These diseases involve the heart or blood vessels and include hypertension, rheumatic heart disease, heart failure, and heart attack. An estimated 17.7 million people in India will die from CVDs representing 31% of all global deaths.

AIM

To perform a systematic review and meta-analysis of the gender prevalence of CVDs in the geriatric population of India.

METHODS

In the present study, we searched databases such as Google Scholar, PubMed and MEDLINE from the year 2003 to 2019 to identify the prevalence of CVDs in the Indian geriatric population. A meta-analysis was conducted using the statistical software R version 3.4.3 and the random effect model was used to determine the pooled estimate of the prevalence of CVDs in the geriatric population of India along with the 95% confidence interval rather than using the fixed effect model. The random effect model takes into consideration the heterogeneity across the various studies.

RESULTS

The prevalence of CVDs in the Indian geriatric population was determined in 6586 male subjects from 32 studies and 8164 female subjects from 32 studies, respectively. The overall prevalence of CVDs in the Indian geriatric population was 36.6% (95%CI: 31.9%-41.3%). In addition, calculation of the various heterogeneity statistics (Cochran’s Q = 3836.85, I² = 98.6%, P < 0.0001) indicated heterogeneity in the prevalence of CVDs in the elderly Indian population in these studies. The prevalence of CVDs in elderly males was 38.0% (95%CI: 33.0%-43.0%) and the prevalence of CVDs in elderly females was 40.9% (95%CI: 35.5%-46.2%).

CONCLUSION

The results indicate that the prevalence of CVDs in the female geriatric population was relatively higher than that in the male geriatric population. Policy makers must take immediate steps to prevent CVDs and improve geriatric health care services in India.

Key Words: Ageing; Cardiovascular diseases; Geriatric; Systematic review; Meta-analysis; Prevalence

Core tip: A systematic review and meta-analysis of the prevalence of cardiovascular diseases in the geriatric population in India from the year 2003 to 2019 revealed that there was a higher prevalence of cardiovascular diseases in the female geriatric population (40.9%) as compared with the male geriatric population (38.0%). In addition, the measures of consistency such as I² and Cochran’s Q suggested heterogeneity between the studies. These findings indicate that health care professionals should take immediate steps to improve geriatric health care services in India.

INTRODUCTION

Ageing is a universal phenomenon associated with deteriorating health status. It is said that nobody grows old merely by living a certain number of years. With the passage of time, certain changes take place in an organism leading to morbidity, disability and even death. Populations around the world are growing old at a high rate with increasing life expectancy. The challenge for health care in the coming years is to ensure quality of life in the geriatric population[1]. The elderly consist of people surpassing the average life span of humans. The contribution of the elderly population to demographic figures is increasing day by day. Increasing problems in health care, psychological, personal and socio-economic factors are associated with the elderly[2]. The growing healthy ageing population is a source of both joy and worry. Joy as people are living longer and healthier lives, and worry about how to respond to a future with a larger older population and their demands and needs[3].

According to the World Health Organization (WHO) Report on Ageing and Health 2018, persons aged 60 years and older are considered elderly. It is suggested that between 2015 and 2050, the proportion of the world’s elderly population will nearly double from 12% to 22%, respectively. By 2050, the world’s population aged 60 years and older is expected to total 2 billion, up from 900 million in 2015[4]. According to the population census 2011, there are nearly 104 million elderly persons in India which constitutes approximately 8% of the total world population. This share of the population is expected to increase from 8% to nearly 12.6% in 2025 and to 19% in 2050[5].

The geriatric population is vulnerable to long-term diseases such as cardiovascular diseases (CVDs), cancer, diabetes, musculoskeletal disorders, depression, arthritis, kidney problems, etc. Due to the increase in life expectancy and modification of lifestyle, CVDs are emerging as one of the major problems in elderly people in India. CVDs generally refer to disorders of the heart and blood vessels and include hypertension, heart attack, cerebrovascular diseases or stroke, peripheral heart disease, rheumatic heart disease, congenital heart disease and cardiomyopathies[6]. The symptoms of heart diseases may vary with respect to disease type. Most CVDs can be prevented by addressing behavioral risk factors such as tobacco use, unhealthy diet and obesity, physical inactivity and harmful use of alcohol, etc. CVDs are a major cause of mortality all over the world as well as in India. According to the WHO Report, an estimated 17.9 million people will die every year from CVDs representing 31% of all deaths worldwide and over 23 million people will die from CVDs by 2030[7]. Our study aims to provide a pooled estimate of the gender prevalence of CVDs in the geriatric population of India.

MATERIALS AND METHODS

Literature search

The search strategy, selection of publications and reporting of results for the review were conducted in accordance to the PRISMA guidelines. All published studies related to the prevalence of CVDs in India over the last 15 years (i.e., from the year 2003 to year 2018) were obtained from MEDLINE/PubMed, Google Scholar and from non-electronic materials such as journals, theses, etc. Internet searches using a combination of keywords such as prevalence, CVDs, hypertension, stroke, elderly, morbidity, etc. were used to identify prevalence studies of CVDs in the Indian geriatric population. The limitations included were: English for the language category and geriatric humans for the study category.

Selection of studies

All identified research papers were independently reviewed by the reviewers. Disagreements between the reviewers were resolved by considering the facts mentioned in the PRISMA checklist.

Inclusion criteria

Studies that met all the following criteria were included in the present analysis: (1) The studies in the mentioned databases with full text, despite the language of the original text; (2) The studies were conducted in the Indian population and were prevalence studies conducted after the year 2002; (3) The study design was cross-sectional; and (4) The age group included in the study was 60 years and older.

Exclusion criteria

Studies which did not satisfy the above criteria were not included in our analysis. We also excluded studies in which age-specific prevalence was not reported. Some studies in which calculated proportions acted as outliers were also excluded from the study to ensure normal distribution of the data.

Data extraction and methodological assessment

Data extraction was performed by the first reviewer and confirmed by the second reviewer, and the whole process was discussed by both reviewers. The eligible studies were further reviewed and information on the author, year of publication, gender, location, total number of subjects, number of cases, type of study, etc. from each included study was extracted and the data were summarized in the form of a table.

Statistical analysis

The meta-analysis was used to combine the results of studies using the statistical software R version 3.4.3. The pooled estimate of prevalence of CVDs in the Indian geriatric population was provided by the random effect model. The random effect model assumes that observed estimates of treatment effect can vary across studies because of real differences in the treatment effect in each study as well as sampling variability (chance). Variation (heterogeneity) across studies must also be considered. The random effect meta-analysis allows for heterogeneity by assuming that underlying effects follow a normal distribution. Cochran’s Q and I² were used to incorporate heterogeneity in the studies and 95% confidence intervals (CIs) are reported in the analysis. Forest plots and funnel plots were used for graphical representation of the meta-analysis. The squares in the forest plot represent the effect estimates of individual studies with their 95%CI of the prevalence of CVDs with the size of squares proportional to the weights assigned to each study in the meta-analysis. The diamond represents the overall result and 95%CI of the random effect meta-analysis. The funnel plots were used to investigate publication bias in the study. The inverted funnel shows that there is no publication bias, whereas the asymmetric funnel plot may suggest publication bias. The dots in the funnel plot represent the studies involved in the meta-analysis. Generally, the studies with larger power are placed towards the top, whereas the lower power studies are placed at the bottom.

RESULTS

Literature review

Figure 1 summarizes the search strategy carried out. Our search strategy identified approximately 720 articles. Of these, 390 duplicate articles were excluded. We also excluded 245 articles after screening the title and abstract. The full text of the remaining 85 articles was screened and 50 articles were further excluded due to lack of inclusion criteria and 35 articles were finally included in the study.

Figure 1 Flowchart showing the selection process of relevant studies.

Study characteristics

We included 35 studies of CVDs in the Indian geriatric population in this meta-analysis. The included articles consisted of 32 studies based on male and female geriatric populations in India. Table 1 shows the characteristics of each study in our analysis after applying the exclusion criteria.

Table 1 The characteristics of each study.

Ref.	Study period	Study design	State	Sampling method	Age group	Sample size
Joshi et al[25]	2001	Cross-sectional	Haryana	Cluster sampling	≥ 60	200
Hazarika et al[9]	2002	Cross-sectional	Assam	Simple random sampling	≥ 60	888
Lena et al[31]	2003	Cross-sectional	Karnataka		≥ 60	213
Prakash et al[2]	2003	Cross-sectional	Rajasthan		≥ 60	300
Kishore et al[29]	2004	Cross-sectional	Uttarakhand		≥ 60	285
Khanam et al[28]	2004	Cross-sectional	Bangladesh	Simple random sampling	≥ 60	452
Bhatia et al[17]	2006	Cross-sectional	Haryana	Stratified random sampling	≥ 65	361
Chandrashekhar et al[1]	2008	Cross-sectional	Karnataka	Systematic random sampling	≥ 60	370
Bharati et al[16]	2008	Cross-sectional	Tamil Nadu	Simple random sampling	≥ 60	225
Bhatt et al[18]	2008	Cross-sectional	Gujarat	Simple random sampling	≥ 60	218
Jacob et al[23]	2010	Cross-sectional	Kerala	Cluster sampling	≥ 60	403
Sharma et al[36]	2010	Cross-sectional	Himachal Pradesh	Multistage simple random sampling	≥ 60	400
Ghosh et al[20]	2011	Cross-sectional	Bihar		≥ 60	431
Kamble et al[26]	2011	Cross-sectional	Maharashtra	Systematic random sampling	≥ 60	494
Qadri et al[33]	2011	Cross-sectional	Haryana	Simple random sampling	≥ 60	660
Banjare et al[14]	2012	Cross-sectional	Odisha	PPS, Systematic sampling	≥ 60	310
Dutta et al[19]	2012	Cross-sectional	Assam	Multistage sampling	≥ 60	370
Banjare et al[13]	2012	Cross-sectional	Odisha	Multistage simple random sampling	≥ 60	310
Barman et al[3]	2013	Cross-sectional	Bihar		≥ 60	160
Shubha et al[11]	2013	Cross-sectional	Karnataka	Simple random sampling	≥ 60	180
Karanth et al[27]	2013	Cross-sectional	Karnataka		≥ 60	500
Alam et al[12]	2014	Cross-sectional	Chhattisgarh	Multistage simple random sampling	≥ 60	640
Bardhan et al[10]	2014	Cross-sectional	Uttar Pradesh	Simple random sampling	≥ 60	980
Bartwal et al[15]	2014	Cross-sectional	Uttarakhand	Systematic sampling	≥ 60	440
Gupta et al[21]	2014	Cross-sectional	Punjab		≥ 60	534
Kumar et al[30]	2014	Cross-sectional	Uttar Pradesh	Stratified random sampling	≥ 60	402
Naushad et al[8]	2014	Cross-sectional	Chhattisgarh	Multistage simple random sampling	≥ 60	640
Jain et al[24]	2014	Cross-sectional	Maharashtra	PPS	≥ 60	600
Noor et al[32]	2014	Cross-sectional	Odisha	Simple random sampling	≥ 60	224
Venkateshkrishna et al[37]	2015	Cross-sectional	Karnataka		≥ 60	1452
Sahu et al[35]	2017	Cross-sectional	Uttar Pradesh		≥ 60	231
Reddy et al[34]	2017	Cross-sectional	Telangana	Systematic random sampling	≥ 60	1265
Gupta et al[22]	2018	Cross-sectional	Haryana	Simple random sampling	≥ 60	300

Meta-analysis

Meta-analysis of the prevalence of CVDs in the Indian geriatric population involved 6586 male subjects from 32 studies. The pooled estimate of prevalence of CVDs in the male geriatric population using the random effect model was 38.0% (95%CI: 33.0%-43.0%) (Table 2). Significant heterogeneity (I² = 95.0% and Cochran’s Q = 621.09, P < 0.0001) among the 32 studies was found (Table 3). The forest plot (Figure 2) and the funnel plot (Figure 3) represent the proportion of subjects affected by CVDs in each study and the pooled estimate of the prevalence of CVDs in the male geriatric population, respectively.

Table 2 Prevalence of cardiovascular diseases in the male Indian geriatric population.

Ref.	n	Cases	Proportion	95%CI	W (Random) (%)
Joshi et al[25]	98	41	0.4184	[0.3195-0.5223]	3.0
Hazarika et al[9]	500	321	0.6420	[0.5982-0.6841]	3.3
Prakash et al[2]	190	84	0.4421	[0.3703-0.5158]	3.2
Kishore et al[29]	177	70	0.3955	[0.3229-0.4716]	3.1
Bhatia et al[17]	152	53	0.3487	[0.2733-0.4301]	3.1
Chandrashekhar et al[1]	181	35	0.1934	[0.1385-0.2585]	3.2
Lena et al[31]	92	53	0.5761	[0.4686-0.6785]	3.0
Bhatt et al[18]	74	19	0.2568	[0.1622-0.3716]	3.0
Bharati et al[16]	92	34	0.3696	[0.2712-0.4766]	3.0
Khanam et al[28]	204	74	0.3627	[0.2968-0.4328]	3.2
Kamble et al[26]	232	50	0.2155	[0.1644-0.2741]	3.2
Sharma et al[36]	196	64	0.3265	[0.2614-0.3970]	3.2
Qadri et al[33]	336	135	0.4018	[0.3489-0.4564]	3.2
Barman et al[3]	88	47	0.5341	[0.4246-0.6412]	2.9
Banjare et al[13]	153	40	0.2614	[0.1938-0.3385]	3.2
Noor et al[32]	132	62	0.4697	[0.3823-0.5585]	3.1
Ghosh et al[20]	196	96	0.4898	[0.4179-0.5620]	3.2
Kumar et al[30]	190	66	0.3474	[0.2799-0.4197]	3.2
Bardhan et al[10]	293	30	0.1024	[0.0702-0.1429]	3.3
Dutta et al[19]	162	57	0.3519	[0.2786-0.4307]	3.1
Banjare et al[14]	153	40	0.2614	[0.1938-0.3385]	3.2
Shubha et al[11]	42	15	0.3571	[0.2155-0.5197]	2.6
Alam et al[12]	267	113	0.4232	[0.3632-0.4849]	3.2
Gupta et al[21]	244	94	0.3852	[0.3239-0.4495]	3.2
Bartwal et al[15]	187	77	0.4118	[0.3405-0.4859]	3.2
Naushad et al[8]	241	113	0.4689	[0.4045-0.5340]	3.2
Jain et al[24]	262	133	0.5076	[0.4454-0.5697]	3.2
Venkateshkrishna et al[37]	530	145	0.2736	[0.2360-0.3137]	3.3
Sahu et al[35]	61	15	0.2459	[0.1446-0.3729]	2.9
Jacob et al[23]	135	77	0.5704	[0.4824-0.6552]	3.1
Reddy et al[34]	594	191	0.3215	[0.2841-0.3608]	3.3
Gupta et al[22]	132	59	0.4470	[0.3604-0.5359]	3.1

Figure 2 Forest plot of the meta-analysis of cardiovascular disease prevalence in the male Indian geriatric population.

Table 3 Tests of heterogeneity among studies of the male Indian geriatric population.

Test of heterogeneity	Prevalence	LCI 95%	HCI 95%
Pooled Statistics	0.38	0.33	0.43
Heterogeneity	4.48	4.02	4.98
I-squared	95.0	93.8	96.0
Cochran's Q	621.09
χ2, P value	0.0001
τ2	0.0197

 

 

 

 

 

 

 

 

 

Figure 3 Funnel plot showing publication bias in the meta-analysis of the prevalence of cardiovascular diseases in the male Indian geriatric population.

Meta-analysis of the prevalence of CVDs in the female geriatric population included 8164 subjects from 32 studies. The pooled estimate of prevalence of CVDs in the female geriatric population using the random effect model was 40.9% (95%CI: 35.5%-46.2%) (Table 4). Significant heterogeneity (I² = 96.4% and Cochran’s Q = 873.0, P < 0.0001) among the 32 studies was found (Table 5). The forest plot (Figure 4) and funnel plot (Figure 5) represent the proportion of subjects affected due to CVDs in each study and the pooled estimate of prevalence of CVDs in the female geriatric population, respectively.

Table 4 Prevalence of cardiovascular diseases in the female Indian geriatric population.

Study name	N	Cases	Proportion	95%CI	%W (Random)
Joshi et al[25]	102	57	0.5588	[0.4571-0.6571]	3.0
Hazarika et al[9]	388	244	0.6289	[0.5787-0.6771]	3.2
Prakash et al[2]	110	60	0.5455	[0.4477-0.6407]	3.0
Kishore et al[29]	108	48	0.4444	[0.3488-0.5432]	3.0
Bhatia et al[17]	209	97	0.4641	[0.3950-0.5342]	3.1
Chandrashekhar et al[1]	189	30	0.1587	[0.1097-0.2188]	3.2
Lena et al[31]	121	73	0.6033	[0.5104-0.6911]	3.0
Bhatt et al[18]	144	56	0.3889	[0.3088-0.4736]	3.1
Bharati et al[16]	122	68	0.5574	[0.4647-0.6472]	3.0
Khanam et al[28]	248	101	0.4073	[0.3455-0.4712]	3.2
Kamble et al[26]	262	69	0.2634	[0.2111-0.3211]	3.2
Sharma et al[36]	204	98	0.4804	[0.4101-0.5513]	3.1
Qadri et al[33]	324	159	0.4907	[0.4351-0.5466]	3.2
Barman et al[3]	72	34	0.4722	[0.3533-0.5935]	2.9
Noor et al[32]	92	46	0.5000	[0.3939-0.6061]	2.9
Ghosh et al[20]	235	82	0.3489	[0.2881-0.4136]	3.2
Kumar et al[30]	212	71	0.3349	[0.2717-0.4028]	3.1
Bardhan et al[10]	197	19	0.0964	[0.0591-0.1465]	3.2
Dutta et al[19]	208	77	0.3702	[0.3044-0.4397]	3.1
Banjare et al[14]	157	20	0.1274	[0.0796-0.1899]	3.2
Shubha et al[11]	138	19	0.1377	[0.0850-0.2066]	3.2
Alam et al[12]	373	207	0.5550	[0.5029-0.6061]	3.2
Karanth et al[27]	500	130	0.2600	[0.2221-0.3008]	3.2
Gupta et al[21]	290	158	0.5448	[0.4856-0.6031]	3.2
Bartwal et al[15]	253	99	0.3913	[0.3308-0.4544]	3.2
Naushad et al[8]	369	207	0.5610	[0.5087-0.6123]	3.2
Jain et al[24]	338	144	0.4260	[0.3727-0.4807]	3.2
Venkateshkrishna et al[37]	922	402	0.4360	[0.4037-0.4687]	3.2
Sahu et al[35]	170	52	0.3059	[0.2376-0.3811]	3.1
Jacob et al[23]	268	124	0.4627	[0.4018-0.5244]	3.2
Reddy et al[34]	671	193	0.2876	[0.2536-0.3235]	3.2
Gupta et al[22]	168	88	0.5238	[0.4455-0.6013]	3.1

Figure 4 Forest plot of the meta-analysis of cardiovascular disease prevalence in the female Indian geriatric population.

Table 5 Tests of heterogeneity among studies of the female Indian geriatric population.

Tests of heterogeneity	Prevalence	LCI 95%	HCI 95%
Pooled Statistics	0.409	0.355	0.462
Heterogeneity	5.31	4.82	5.84
I-squared	96.4	95.7	97.1
Cochran's Q	873.0
χ2, P value	0.0001
τ2	0.0229

Figure 5 Funnel plot showing publication bias in the meta-analysis of prevalence of cardiovascular diseases in the female Indian geriatric population.

DISCUSSION

CVD in the geriatric population is emerging as a major problem in India. In the present analysis, we attempted to provide an estimate of the pooled prevalence of CVDs among the Indian geriatric population along with the 95%CI. The estimated overall prevalence of CVDs in the Indian geriatric population was 36.6% (95%CI: 31.9%-41.3%).

The estimated prevalence of CVDs in the male geriatric population was 38.0% (95%CI: 33.0%-43.0%) which was less than the prevalence (46.9%) reported by Naushad et al[8] in Raipur, Chhattisgarh, India. The highest prevalence (64.2%) of CVDs in the male Indian geriatric population was observed by Hazarika et al[9]. Similar findings were reported by Lena et al[31] and Barman et al[3] in their respective studies. Rapid urbanization, lifestyle changes, dietary changes and increased life expectancy were factors attributed to this rising trend. The lowest prevalence (10.2%) of CVDs in the male Indian geriatric population was observed by Bardhan et al[10]. Chandrashekhar et al[1] also reported similar findings in their study in which the prevalence of CVDs in the geriatric population was 19.3%.

The estimated overall prevalence of CVDs in the female geriatric Indian population was 40.9% (95%CI: 35.5%-46.2%) which is higher than the prevalence (35.7%) reported by Shubha et al[11]. The highest prevalence (62.9%) of CVDs in the female Indian geriatric population was observed by Hazarika et al[9] Bharati et al[16], Joshi et al[25], Naushad et al[8] and Alam et al[12] who reported a similar high prevalence of CVDs in female subjects. The lowest prevalence (9.6%) was observed by Bardhan et al[10]. Studies by Banjare et al[14] and Shubha et al[11] also showed a similar pattern of low prevalence of CVDs in elderly female subjects.

The current statistics for CVDs show that almost 80% of premature heart attacks and strokes are preventable. The government should promote awareness of these diseases through mass media and other means and provide information on diet and physical health.

Conclusion

The present systematic review and meta-analysis demonstrated that the prevalence of CVDs in the male and female geriatric population in India was estimated to be 38.0% and 40.9%, respectively. These considerable prevalence rates require supportive interventions for the prevention and early diagnosis of CVDs in the geriatric population. It is imperative to provide geriatric health services at primary health centers in addition to increasing the awareness among elderly people of health, diseases and health care facilities in order to improve their quality of life.

Limitations

Given the limited amount of data available on the prevalence of CVDs in the geriatric population in India; studies conducted in subjects with different bio-social characteristics have been included in the meta-analysis. Also, the pooled estimate of prevalence of CVDs in the Indian geriatric population does not provide an overview of the problem of CVDs.

Implications for future research

The prevalence of CVDs in the Indian geriatric population presents a formidable challenge to the Indian health system. In future, researchers should identify the location- and age-specific pooled estimate of CVDs in the geriatric population of India. Other areas of research should include determination of the pooled prevalence of other morbidities such as diabetes, arthritis, etc. in the geriatric population.

ARTICLE HIGHLIGHTS

Research background

India is in the phase of demographic transition. From the 2011 census, there are 104 million elderly persons in India as compared with 57 million elderly persons in 1991. The elderly are exposed to several morbidities such as cardiovascular diseases (CVDs), diabetes, hypertension, etc. CVDs have become the leading cause of mortality in India. An estimated 17.7 million people in India will die from CVDs representing 31% of all global deaths. Therefore, a precise estimate of the prevalence of CVDs in elderly males and females in India is required to assess the magnitude of the problem which needs to be addressed.

Research motivation

Very few studies are available on the prevalence of CVDs in India. The gender pooled estimate of CVDs from various studies conducted in different regions of the country can aid the development of preventive strategies. The difference in the prevalence of CVDs among male and female geriatric subjects was studied to provide an insight into the type of preventive and promotional services required.

Research objectives

The main objective of the study was to provide the gender pooled estimate of the prevalence of CVDs in the geriatric population of India. The large increase in the prevalence of CVDs in the Indian geriatric population has prompted health managers to take immediate steps for the prevention and early detection of these diseases. In addition, researchers are attempting to identify the location- and age-specific pooled estimate of the prevalence of CVDs. The main application of this meta-analysis is to consolidate the available data to determine the burden of CVDs in India.

Research methods

Secondary data related to the prevalence of CVDs in the geriatric population were collected from published research papers and then analyzed using meta-analysis in R software. A meta-analysis integrates the quantitative findings from separate studies and provides a numerical estimate of the overall effect of interest.

Research results

The overall prevalence of CVDs in the Indian geriatric population was estimated to be 36.6% (95%CI: 31.9%-41.3%). The prevalence of CVDs in the male geriatric population was 38.0%, whereas the prevalence of CVDs in the elderly female population was 40.9%.

Research conclusions

The pooled prevalence of CVDs in the female Indian geriatric population was greater than the pooled prevalence of CVDs in the male Indian population. These findings will help policy makers to take immediate steps to provide geriatric health care services in India. It is possible that an insight into the magnitude of the problem of CVDs can help to shape preventive programs for CVDs.

Research perspective

Future studies should be conducted to determine the pooled estimate of CVDs in the geriatric population in both rural and urban areas, and in different age-groups. Subgroup analysis in meta-regression can be used to calculate these values.