This Article Figures Only Full Text (PDF) Online Appendix Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Prabhakaran, D. Articles by Reddy, K. S. Search for Related Content PubMed Articles by Prabhakaran, D. Articles by Reddy, K. S. Related Collections Related Article

FROM AROUND THE WORLD: FOCUS ON INDIA

Impact of a Worksite Intervention Program on Cardiovascular Risk Factors

A Demonstration Project in an Indian Industrial Population

Dorairaj Prabhakaran, MD, DM, MSc^_*, Panniyammakal Jeemon, MPH^_*, Shifalika Goenka, MBBS, PhD^_*, Ramakrishnan Lakshmy, MD, K.R. Thankappan, MD, MPH, Faruq Ahmed, MD, Prashant P. Joshi, MD^||, B.V. Murali Mohan, MD^¶, Ramanathan Meera, MBBS, MPH^#, Mohas S. Das, MD, DM^_**, Ramesh C. Ahuja, MD, DM, Ram Kirti Saran, MD, DM, Vivek Chaturvedi, MD, DM and K. Srinath Reddy, MD, DM, MSc^,,_*

^_* Initiative for Cardiovascular Health Research in the Developing Countries, New Delhi, India
All India Institute of Medical Sciences, New Delhi, India
Sree Chitra Tirunal Institute of Medical Sciences and Technology, Trivandrum, Kerala, India
Khajabandanawaz Institute of Medical Sciences, Gulberga, Karnataka, India
^|| Government Medical College, Nagpur, India
^¶ Ambedkar Medical College, Bangalore, India
^# PSG Medical College, Coimbatore, India
^_** Nizam's Institute of Medical Sciences, Hyderabad, India
King George Medical College, Lucknow, India
Public Health Foundation of India, New Delhi, India

^_* Reprint requests and correspondence: Dr. K. Srinath Reddy, Public Health Foundation of India, PHD House, Siri Fort Institutional Area, New Delhi 110016, India (Email: ksreddy{at}ccdcindia.org).

Key Words: worksite interventions • cardiovascular disease • prevention and control • India

Although reasonable evidence exists for the beneficial role of risk factor reduction in decreasing CVD risk among individuals at high risk, primary or primordial prevention programs that use population-based approaches have yielded equivocal results (8,9). For example, a meta-analysis of all population-based studies conducted largely in developed countries has suggested that health promotion (involving health education, mass media, and community organization) does not reduce mortality significantly but leads to small yet potentially beneficial reduction in risk factor levels (10). Several reasons have been attributed to this equivocal result of health promotion. These include shorter duration of intervention, improper design to evaluate the benefits, contamination (adoption of components of health intervention by the control community), and a declining trend of CVD in developed countries during the intervention period. However, by contrast, in developing countries the current prevailing secular trend seems to be a rapidly increasing burden of CVD and its risk factors. Therefore it is likely that a community-based approach may show the desired results of reducing CVD risk factors in developing country settings. For example, a primary prevention and health promotion initiative in Mauritius showed a pronounced decrease in the population level total cholesterol concentrations after 5 years of the intervention program (11).

India is experiencing an accelerated epidemiological transition with a consequent increase in the burden of CVD risk factors both in community-based studies and in industrial populations (12–16). Given this background, we hypothesized that a comprehensive CVD risk factor reduction program comprising of a multipronged strategy of health promotion, high-risk primary prevention, and policy level or environmental changes and using existing infrastructure in the participating industries would yield substantial reductions in CVD risk factors. We outline the methods of developing such a comprehensive CVD prevention and health promotion program, present the results of this program, and discuss their implications.

	Methods

Top Methods Results Discussion Conclusions Appendix References

 
The design of the overall study is presented in a flowchart (Fig. 1). The baseline CVD risk factor survey covered 10 different industrial sites representing multiple regions of India. The detailed methodology and results of the survey are published elsewhere (16,17). Briefly, all of the employees and their family members between the ages of 10 and 69 years were eligible to be included in the survey. Age group (in deciles) and a sex-stratified multistage random sampling technique was used for the baseline survey. At each participating center, detailed data were obtained from 800 randomly selected employees and their eligible family members at baseline. The study involved collection of data related to the demographic profile, individual characteristics related to major risk factors of CVD, past medical history, clinical and anthropometric profile, and biochemical parameters. Strict quality control measures were taken to ensure the accuracy, completeness, and comparability of blood pressure, anthropometric, and biochemical measurements across the 10 study sites. The details of quality control measures are published elsewhere (16). Complete risk factor details including the biochemical variables at the baseline were available for 10,442 subjects, and the response rate was 89.7%.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 1 Study Flow Chart Overview of study methodology.

Interventions.   A multicomponent, multilevel, and multimethod intervention was implemented by trained, locally stationed, project health care personnel in the participating industries, using the industry as a setting, as a target, and as an agent as well as a resource, over 4 consecutive years. A population-based approach was the mainstay of the intervention, which was augmented by high-risk and policy change/environmental approaches.

Conceptual framework and objectives.   The socioecological model, the social cognitive theory of behavior change, and the social learning theory provided the conceptual framework for the health promotion component. The primary objectives of the intervention were: to increase the consumption of locally available fruits and vegetables and to move toward a healthy diet on the whole (involving use of higher fiber and decreased salt and oils with the optimal mix of polyunsaturated and monounsaturated fats, and so on). Other interventions included incorporation of physical activity in daily living, avoidance of tobacco and its products, and maintenance of a healthy body weight. In addition, we sensitized individuals on the need for recognizing and treating high blood pressure and diabetes. Contextual and cultural understanding and formative feedback from the various industries significantly contributed to the thematic content of the intervention materials.

Multiple components of the intervention.   These were posters and banners, handouts, booklets, and real-time videos that were translated into 7 Indian languages for the culturally and linguistically diverse population. These were conceived, designed, and developed based on an understanding of the various scientific theories of behavior change, recent scientific literature on cardiovascular risk factors and interventions, cultural and contextual formative inputs, and contemporary techniques used for creating, developing, and designing effective tools. Some of the materials used during the intervention period are available at: http://www.whoindia.org/EN/Section20/Section385_1540.htm.

Multilevel interventions.   The interventions sought to influence behaviors at the individual level, interpersonal level (family and workplace-related peers), and environmental/macro level (social norms at the worksite and home).

Methods of intervention.   We used several methods of intervention. These included direct one-on-one interactions between the trained health project personnel and the employees and their families, and population-based strategies through the use of posters, banners at strategic locations in the industry, handouts, booklets, and different video films shown on the internal cable network for individual and mass sensitization. Additionally, dynamic group interactions, health melas (health display), and motivational sessions were conducted by the investigators and the locally stationed project personnel from time to time. Motivation and empowerment were 2 crucial aspects of our health intervention, and motivated families, groups, and individuals acted as propagators of our messages. Consequently, an array of policy and environmental changes were initiated by a motivated management and by employees themselves. For example, the canteen menu was modified to include salads and the frequency of fried items served per week was reduced; pickles and papads (a thin Indian wafer), which are high in salt, were voluntarily withdrawn or restricted; and on-site use of tobacco and its products were banned. On certain days fruits were provided as an alternative to Indian deserts. In addition, individuals with established risk factors were encouraged to be compliant with their treatment and were referred to a health care facility for further risk management. Risk stratification and treatment guidelines for diabetes and hypertension were prepared for on-site physicians, and targets were provided for risk factor reduction. Special individual and group counseling sessions on diet, tobacco use, and physical activity were also conducted for those who had established risk factors. The participants were allowed to contact the trained project staff for individual counseling sessions as and when they wanted.

Interventions in the control site.   In the control site, no project staff were available during the intervention phase and none of the above activities were carried out. However, all individuals with established CVD risk factors were referred to the industry-managed clinic for further follow-up after the baseline survey. The industry management was free to organize any health promotion activities during the intervention phase. They had organized 3 human immunodeficiency virus/acquired immunodeficiency syndrome awareness programs and a health talk on the ill effects of tobacco during the intervention phase. The site management also banned tobacco use inside the premises of the industry.

Data analysis.   Initially, the general characteristics of the 2 populations were compared. An independent t test was used for comparison of means, and a chi-square test was used for comparison of proportions. The percentage change in risk factor levels and the 95% confidence interval (CI) of this change (CI of the mean difference of the 2 independent populations was calculated and converted into percentage change from the baseline value) in the intervention group and the control group were compared independently. A separate analysis was carried out in a subset of the sample population (those who attended both the baseline and the final surveys), and the mean difference in CVD risk factor levels in both groups were compared using a mixed linear regression model. Each dependent variable was adjusted for age, sex, education status at baseline, baseline mean level of the same variable, and body mass index. The Duncan t test p value was also calculated. A physical activity scoring system was developed based on self-reported energy-consuming activities during work, at home, while traveling, and at leisure time. Activities were graded on an ordinal scale with scores ranging from 1 to 4, and the cumulative physical activity (score ranging from 4 to 16) was calculated by adding the scores during work, at home, while traveling, and at leisure time. The median cumulative score before and after intervention was compared using nonparametric tests. All analyses were carried out using the statistical software SPSS for Windows (version 13.1, SPSS Inc., Chicago, Illinois).

	Results

Top Methods Results Discussion Conclusions Appendix References

 
Comparison of baseline characteristics of the study population.   The baseline characteristics of the study population stratified by the intervention group are given in Table 1. The mean age at baseline in the intervention group was 40.8 (SD 10.8) years, compared with 38.6 (SD 11.7) years in the control group. The majority of participants were male, both in the intervention group and in the control group (58.7% and 58.1%, respectively). More than three-fourths of the study population had education above primary school level (79.0% in the intervention group and 85.7% in the control group).

View larger version (6K): [in this window] [in a new window] [Download PPT slide]   Figure 2 Comparison of Percentage Change in Mean Risk Factors in Intervention and Control Site Before-and-after differences in cardiovascular disease risk factors. 1 = weight; 2 = waist circumference; 3 = systolic blood pressure; 4 = diastolic blood pressure; 5 = plasma glucose; 6 = total cholesterol; 7 = high-density lipoprotein cholesterol; 8 = serum triglycerides. The horizontal line for each variable represents the point estimate, and the ends of the vertical line represent 95% confidence interval of the point estimate.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 3 Comparison of Age-Stratified Mean Levels of SBP and TC (2001 to 2002 and 2006 to 2007) Before-and-after differences in mean systolic blood pressure (SBP) and total cholesterol (TC) across age groups.

View larger version (26K): [in this window] [in a new window] [Download PPT slide]   Figure 4 Prevalence of Tobacco Use, Extra Salt Consumption, Fruit Consumption, and Median Physical Activity Score in Intervention and Control Communities Self-reported changes in health-related behaviors.

View larger version (21K): [in this window] [in a new window] [Download PPT slide]   Figure 5 Proportion of Individuals Above the Framingham 10-Year Cardiovascular Disease Risk of 10% Before-and-after differences in cumulative cardiovascular risk.

View larger version (6K): [in this window] [in a new window] [Download PPT slide]   Figure 6 Percentage Mean Difference in Change in Cardiovascular Disease Risk Factors in Cohort Analysis Adjusted mean differences in change in cardiovascular disease risk factors after intervention. Mixed linear model adjusted for age, educational status, sex, baseline body mass index, and baseline mean of the same variable. The horizontal line for each variable represents the point estimate, and the ends of the vertical line represent the 95% confidence interval of the point estimate. DBP = diastolic blood pressure; HDLc = high-density lipoprotein cholesterol; P Glucose = plasma glucose; Tg = triglycerides; WC = waist circumference; Wt = weight; other abbreviations as in Figure 3.

	Discussion

Top Methods Results Discussion Conclusions Appendix References

In this demonstration project, health promotion education efforts were complimented with environmental/policy changes and a high-risk approach. The comprehensive approach of targeting multiple risk factors through different strategies was successful in achieving a significant reduction in SBP and DBP in the intervention group. Although the blood pressure effects of lifestyle modifications in general are modest, they seem to be additive when multiple measures are considered in the intervention package (19). The significant changes in related behaviors (reduction in consumption of extra salt, increase in physical activity levels and in fruits and vegetable consumption, and decrease in use of tobacco products) as documented in the intervention population could have contributed to the significant reduction in blood pressure. The results have significant implications, because more than a 2-fold difference in the stroke death rate and nearly a 2-fold difference in the death rates from ischemic heart disease and from other vascular causes are observed with each difference of 20 mm Hg usual SBP (or, approximately equivalently, 10 mm Hg usual DBP) in the middle-age population over an average period of 10 years (20). Furthermore, even a smaller 10-mm Hg SBP lowering with a range of drug therapies in a large series of clinical trials estimates approximately a 20% to 25% reduction in major cardiovascular events (21).

The mean differences in change in serum cholesterol levels before and after intervention, in the intervention group, to the magnitude of a 10.7 mg/dl decrease in total cholesterol and a 4.6 mg/dl increase in HDL cholesterol also have significant implications at the population level. A 10% reduction in serum total cholesterol levels is associated with a decrease in risk of ischemic heart disease of about 50% at age 40 years (22). A reduction in serum cholesterol to the extent of 10% is realistic on a community basis only when individuals change their dietary behaviors in conjunction with family, friends, and workmates and the dietary change is perceived more positively (22).

Interventions that reduce glucose levels at the population level will reduce the incidence of new-onset diabetes and the overall risk of developing a cardiovascular event. The observed reduction in mean blood sugar levels in the intervention group was significant, and the resultant relative CVD risk reduction in the target community would be potentially significant.

The observed reduction in prevalence of tobacco use in the intervention group was significant. This would be expected to have caused an increase in weight and in the prevalence of obesity variables as reported in other studies (23,24). However, the weight reduction that we have achieved through the intervention strategy was apparently able to more than nullify the changes in weight, if any, caused by a decline in tobacco prevalence. The modest increase in tobacco prevalence seen in the control group was consistent with the trend in increase in prevalence of tobacco use in the general population as reported in nationally representative periodic survey analysis (25).

The cumulative effect of reduction in blood pressure, serum total cholesterol, and smoking prevalence and increase in HDL cholesterol levels results in a significant reduction in individuals with a 10-year Framingham risk score of 10 in the intervention community (34.1% at baseline to 26.8% at the final survey). This will prevent approximately 871 cardiovascular events in the target community (73 events per 1,000 population) over the next 10-year period, assuming no change in the risk score without the interventions. However, this may be an underestimate because the proportion of individuals with a 10-year Framingham risk score increased significantly in the control group (25.4% at baseline to 34.7% at the final survey) during the same period.

Rose (26) hypothesized that behavioral changes through population-wide strategies for reducing the incidence of CVD and diabetes would yield significantly more public health benefits than a high-risk approach alone. However, such population-based strategies for control of CVD and diabetes yielded largely inconclusive results (8). The mixed results produced skepticism among the scientific community on the effectiveness of these programs. The reasons for such equivocal results, we believe, are several. For example, most of these studies were conducted in the 1970s and 1980s in developed countries that were already observing a declining CVD incidence. Further, the periods of interventions and follow-up were short. Other potential reasons include contamination of the control population, baseline differences in the comparison group, quasiexperimental designs, and ignoring the effects of design in the analysis. Additionally, the "dose" of the intervention was not adequate enough to achieve the desirable changes. We believe that our program was successful because of the synergy of combining community intervention, high-risk approaches, and policy-level changes. Similarly, a worksite intervention program combining a population strategy and a high-risk strategy in the Capital Iron and Steel Company of Beijing showed a significant reduction in stroke mortality and morbidity (27).

In India, which has a large population at risk for CVD, even small reductions in the population risk factor profiles should result in larger reductions of morbidity and mortality caused by CVD, thereby saving millions from disability and expenditure on curative measures. Given that the human capital in the organized sector itself is huge (almost 30 million workforce in the country), this program if demonstrated effectively through a randomized trial has the potential to make a national impact. India's vast population size and social, economic, and cultural diversities underscore the importance of developing a culturally relevant and context-specific health promotion strategy tailored to the needs of the population. Further, the total direct cost of this program was approximately U.S. $7.3 per person per year in research mode, and will be even lower when carried out as a larger program. Given the fact that India is expected to lose 237 billion international dollars from 2005 to 2015, an economic loss attributable to CVD, we believe such a low-cost program would be extremely useful (28).

Study limitations.   The nonrandomized design and the results based on individuals as the unit of analysis are the major limitations of this study. However, comparison of the intervention group with a comparable industrial population without any interventions improved the quality of the results. Although nonrandomized cluster design studies are invariably contaminated by the secular trends, the data trend in the control group in this study mirrors the prevailing secular trend in increasing CVD risk factors in the general population in India (11,12,25). Therefore, it would be hard to argue that the outcome in the intervention group is attributable to selection bias, particularly the selection-maturation threat (a threat to the internal validity of the study that combines selection bias with a maturation threat, i.e., differential rates of normal growth between pre- and post-test for the groups). A relatively small number of clusters prevents us from adjusting for the variation in considering individuals as the unit of analysis. The mixed linear model adjusting for the heterogeneity in changes in outcome variables across the clusters is considered to be the best model available in similar situations for data analysis (29). Several measures, such as smoking, salt consumption, and physical activity, were self-reported and not subject to verification. However, the objective measurements, such as weight, SBP, blood sugars, and lipids, mirrored the self-reported variables, and there is therefore an overall consistency in the trends.

	Conclusions

Top Methods Results Discussion Conclusions Appendix References

 
The data suggest that a worksite approach in health promotion programs on cardiovascular risk factors can be implemented and can have a positive impact on intermediate CVD outcomes in developing countries. A comprehensive approach targeting multiple risk factors and a systematic, randomized, controlled design with adequate power to detect the impact of the expected changes on hard cardiovascular end points is warranted.

	Appendix

Top Methods Results Discussion Conclusions Appendix References

 
For a complete list of acknowledgments and investigators, please see the online version of this article.

	Footnotes

 
Please note: This study was funded by the World Health Organization, Country Office, Nirman Bhavan, New Delhi, and the Ministry of Health and Family Welfare, Government of India.

	References

Top Methods Results Discussion Conclusions Appendix References

 
1. Mathers CD, Lopez A, Stein D, et al. Deaths and Disease Burden by Cause: Global Burden of Disease Estimates for 2001 by World Bank Country GroupsDisease Control Priorities Project working paper no. 18. Fogarty International Center.Bethesda, MD: National Institutes of Health; 2001.

2. World Health Organization The World Health Report 2004: Changing HistoryGeneva: World Health Organization; 2004.

3. Murray CJL, Lopez AD. Global Health Statistics. Global Burden of Disease and Injury Series. Boston, MA: Harvard School of Public Health; 1996.

4. World Health Organization The World Health Report 2002: Reducing Risks, Promoting LifeGeneva: World Health Organization; 2002.

5. Reddy KS. Cardiovascular disease in non-Western countries N Engl J Med 2004;350:2438-2440.[Free Full Text]

6. Rose G. International trends in cardiovascular disease—implications for prevention and treatment Aust N Z J Med 1984;14:375-380.[Web of Science][Medline]

7. Jamison TD, Breman GJ, Measham RA, et al. Disease control priorities in developing countriesIn: Rodgers A, Lawes CMM, Gaziano T, et al. editors. The Growing Burden of Risk From High Blood Pressure, Cholesterol and Bodyweight. New York, NY: The World Bank and Oxford University Press; 2006. pp. 851-869.

8. Deborah E, Sellers L, Crawford KB, McKinlay JB. Understanding the variability in the effectiveness of community heart health programs: a meta-analysis Soc Sci Med 1997;44:1325-1339.[CrossRef][Web of Science][Medline]

9. Manuel DG, Lim J, Tanuseputro P, et al. Revisiting Rose: strategies for reducing coronary heart disease BMJ 2006;332:659-662.[Free Full Text]

10. Lim SS, Gaziano TA, Gakidou E, et al. Prevention of cardiovascular disease in high-risk individuals in low-income and middle-income countries: health effects and costs Lancet 2007;370:2054-2062.[CrossRef][Web of Science][Medline]

11. Dowse GK, Gareeboo H, Alberti KG, et al. Mauritius Non-Communicable Disease Study Group Changes in population cholesterol concentrations and other cardiovascular risk factor levels after five years of the non-communicable disease intervention program in Mauritius BMJ 1995;311:1255-1259.[Abstract/Free Full Text]

12. Gupta R, Gupta VP. Meta analysis of coronary heart disease prevalence in India Indian Heart J 1996;48:241-245.[Medline]

13. Gupta R, Gupta VP, Sarna M, Prakash H, Rastogi S, Gupta KD. Serial epidemiological surveys in an urban Indian population demonstrate increasing coronary risk factors among the lower socioeconomic status J Assoc Physicians India 2003;51:470-477.[Medline]

14. Mohan V, Shanthirani S, Deepa R, Premalatha G, Sastry NG, Saroja R. Chennai Urban Population Study (CUPS No4). Intra-urban differences in the prevalence of the metabolic syndrome in southern India—the Chennai Urban Population Study (CUPS No. 4). Diabet Med 2001;18:280-287.[CrossRef][Web of Science][Medline]

15. Prabhakaran D, Chaturvedi V, Shah P, et al. Differences in the prevalence of metabolic syndrome in urban and rural India: a problem of urbanization Chronic Illn 2007;3:8-19.[Free Full Text]

16. Reddy KS, Prabhakaran D, Chaturvedi V, et al. Methods for establishing a surveillance system for cardiovascular diseases in Indian industrial populations Bull World Health Organ 2006;84:461-469.[CrossRef][Web of Science][Medline]

17. Reddy KS, Prabhakaran D, Jeemon P, et al. Educational status and cardiovascular risk profile in Indians Proc Natl Acad Sci U S A 2007;104:16263-16268.[Abstract/Free Full Text]

18. Wilson WF, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories Circulation 1998;97:1837-1847.[Abstract/Free Full Text]

19. The Trials of Hypertension Prevention Collaborative Research Group Effects of weight loss and sodium reduction intervention on blood pressure and hypertension incidence in overweight people with high-normal blood pressure Arch Intern Med 1997;157:657-667.[Abstract/Free Full Text]

20. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R, Prospective studies collaboration Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies Lancet 2002;360:1903-1913.[CrossRef][Web of Science][Medline]

21. Blood Pressure Lowering Treatment Trialists' Collaboration Effects of different blood-pressure lowering regimens on major cardiovascular events: results of prospectively-designed overviews of randomized trials Lancet 2003;362:1527-1535.[CrossRef][Web of Science][Medline]

22. Yusuf S, Cairns AJ, Camm AJ, Fallen EL, Gersh BJ. Evidence based cardiology. Lipids and cardiovascular disease. BMJ 2003:121-129.

23. Froom P, Kristal-Boneh E, Melamed S, Gofer D, Benbassat J, Ribak J. Smoking cessation and body mass index of occupationally active men: the Israeli CORDIS study Am J Public Health 1999;89:718-722.[Abstract/Free Full Text]

24. Williamson DF, Madans J, Anda RF, Kleinman JC, Giovino GA, Byers T. Smoking cessation and severity of weight gain in a national cohort N Engl J Med 1991;324:739-745.[Abstract]

25. Thankappan KR, Mini GK. Case-control study of smoking and death in India N Engl J Med 2008;358:2842-2843.[Free Full Text]

26. Rose G. Sick individuals and sick populations Int J Epidemiol 1985;14:32-38.[Abstract/Free Full Text]

27. Chen J, Wu X, Gu D. Hypertension and cardiovascular diseases intervention in the capital steel and iron company and Beijing Fangshan community Obes Rev 2008:142-145;9 Suppl 1:.

28. World Health Organization Preventing Chronic Disease: A Vital InvestmentGeneva: World Health Organization; 2005.

29. Murray DM, Varnell SP, Blitstein JL. Design and analysis of group-randomized trials: a review of recent methodological developments Am J Public Health 2004;94:423-432.[Abstract/Free Full Text]

Related Article

Inside This Issue
J. Am. Coll. Cardiol. 2009 53: A24. [Full Text] [PDF]

This article has been cited by other articles:

				D. Prabhakaran and A. Roy Commentary: Societal influences on cardiovascular disease: time to assess and act Int. J. Epidemiol., September 23, 2009; (2009) dyp295v1. [Full Text] [PDF]

This Article Figures Only Full Text (PDF) Online Appendix Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Prabhakaran, D. Articles by Reddy, K. S. Search for Related Content PubMed Articles by Prabhakaran, D. Articles by Reddy, K. S. Related Collections Related Article