T H E
Communication
Computer Use, Internet Access, and Online Health Searching Among Harlem Adults Alwyn T. Cohall, MD; Andrea Nye, MPH, MBA; Joyce Moon-Howard, DrPH; Rita Kukafka, DrPH; Bonnie Dye, MPH; Roger D. Vaughan, DrPH, MS; Mary E. Northridge, PhD, MPH
Abstract Purpose. Computer use, Internet access, and online searching for health information were assessed toward enhancing Internet use for health promotion. Design. Cross-sectional random digit dial landline phone survey. Setting. Eight zip codes that comprised Central Harlem/Hamilton Heights and East Harlem in New York City. Subjects. Adults 18 years and older (TV = 646). Measures. Demographic characteristics, computer use, Internet access, and online searching for health information. Analysis. Frequencies for categorical variables and means and standard deviations for continuous variables were calculated and compared with analogous findings reported in national surveys from similar time periods. Results. Among Harlem adults, ever computer use and current Internet use were 77% and 52%, respectively. High-speed home Internet connections were somewhat lower for Harlem adults than for U.S. adults overall (43% vs: 68%). Current Internet users in Harlem were more likely to be younger, white vs. black or Hispanic, belter educated, and in better self-reported health than non-current users (p < .01). Of those luho reported searching online for health information, 74 % sought information on medical problems and thought that information found on the Internet affected the way they eat (47%) or exercise (44%). Conclusions. Many Harlem adults currently use the Internet to search for health information. High-speed connections and culturally relevant materials may facilitate health information searching for underserved groups. (Am J Health Promot 2011;25[5]:325-333.)
Key Words: Health Promotion, Health Education, Communication, Media, Urban Health, Race/Ethnicity, Prevention Research. Manuscript format: research; Research purpose: descriptive; Study design: quasi-experimental; Outcome measure: behavioral; Setting: local community; Health focus: fitness/physical activity, medical self care, nutrition; Strategy: education; Target population age: adults; Target population circumstances: race/ethnicity
Authors are with the Harlem Health Promotion Center, Columbia University Mailman School of Public Health, New York, New York. Send reprint requests to Mary E. Northridge, PhD, MPH, Columbia University Mailman School of Public Health, Department of Sociomedical Sciences and the Harlem Health Promotion Center, 722 W. 168th Street, Room 525, New York, NY 10032;
[email protected]. This manuscript was submitted March 25, 2009; revisions were requested September 4 and September 16, 2009; the manuscript was accepted for publication September 22, 2009. Copyright© 2011 by American Joumal of Health Promotion, Inc. 0890-1171/11/S5.00 + 0 DOI: 10.4278/ajhp.090325-QlJAN-121
American Journal of Health Promotion
PURPOSE The Internet has brought about an opportunity for consumers to bypass the formal health sector and go directly to a source for information and services delivered on demand.' This shift in the availability of health information has precipitated a change in the way patients approach their own health and health care.^ For instance, more than 80% of Americans with Internet access have sought health information online, and surveys indicate that the Internet is viewed by consumers as an important resource to help modify their health behaviors and to find and evaluate health care providers, medications, and therapeutic regimens.^ Thus, the Internet has become an important adjunct to traditional sources of health information for many Americans and a tool enabling consumers to be proactive rather than passive regarding their health.""* Further, the Internet holds promise in assisting marginalized groups to obtain much needed health resources and social support.^^ Nonetheless, disparities in access still exist between people who have access to the Internet and those who do not, often referred to as the "digital divide."^ In general, younger, more affiuent, and more highly educated whites tend to be better connected than older, poorer, and less well-educated members of racial and ethnic minorities. Finally, only about one-half of Americans with chronic health conditions use the Internet, as compared with 74% of their healthy counterparts.'" The profiles of individuals with reduced Internet access mirrors those
May/June 2011, Vol. 25, No. 5
325
of individuals wbo bave been historically underserved by tbe U.S. bealth care delivery system.^ In particular, lower-income Americans, particularly racial and etbnic minorities, suffer disproportionately from cbronic bealtb conditions, are less likely to receive routine medical services, and experience bigber rates of morbidity and mortality tban the general U.S. population." Without directed intervention, development, and dissemination of electronic health care initiatives to marginalized and underserved populations, the digital divide may widen because tbose wbo are better off are often tbe first to implement new ideas and tecbnologies. Indeed, differential uptake among consumers by socioeconomic status bas been documented upon tbe introduction into communities of information obtained online.''^"'"'' Tbe U.S. Department of Healtb and Human Services bas recognized tbe potential of tbe Internet as a mecbanism for improving bealtb and consequently underscored tbe importance of devoted attention to underserved groups. Specifically, Healthy People 2010 objective 11-1 aims to increase tbe proportion of botisebolds witb Internet access at home from a baseline of 26% measured in 1998 to a targeted goal of 80% in 2010.'** In response, studies focused primarily on Internet access and utilization patterns of underserved groups, such as black and Hispanic adults, regarding bealtb information and social support bave been conducted to monitor progress toward tbis targeted goal.''"'^"^® We bypotbesized that better understanding of Internet access issues and health information-seeking bebaviors will belp bring about solutions to online searcbing for bealtb information in tbe form of culturally and socially appropriate Internet interventions aimed at improving bealtb outcomes for adults as tbey age. To provide a platform for planned researcb, education, and service delivery activities of tbe Harlem Healtb Promotion Center (HHPC), we sougbt to examine tbe demograpbic cbaracteristics, bealth status, extent of computer use and Internet access, and online health information-seeking bebaviors of a sample of Harlem adults.
326
American Journal of Health Promotion
METHODS Design A random digit dial (RDD) landline pbone survey was conducted in tbe eigbt zip codes that comprise the Central Harlem/Hamilton Heights (10026, 10027, 10030, 10031, 10037, and 10039) and East Harlem (10029 and 10035) neigbborboods of New York City. (Tbe survey instrument is available upon request from tbe autbors.) Appropriate Columbia University Institutional Review Board and Healtb Insurance Portability and Accountability Act safeguards were followed. Sample Tbe sample size was selected to produce a margin of error of ±4%. Upon agreement to participate in tbe survey, one adult 18 years or older was randomly selected from tbe bousehold to complete the interview. Fielding began in January 2006 and ended in April 2006. An independent company, Abt SRBI (bttp;//www.srbi.com/bome.btml), was contracted to select an RDD sample of 675 adults 18 years or older residing in Harlem for an individual pbone survey. Researcbers held training sessions with interviewers to ensure comprehension of eacb item and to answer any questions from tbe interviewers. Ten telepbone pilot interviews were conducted by Abt SRBI and monitored by HHPC researcb staff. To improve participation rates, direct marketing approacbes were used to familiarize people witb tbe study design and purpose. Two weeks prior to tbe beginning of tbe pbone survey, a postcard was mailed to every household in Harlem that comprised the sampling frame for tbe RDD selection process to alert residents to tbe survey and explain tbe purpose of tbe study. In addition, fiyers were placed in stores and commtmity-based organizations in tbe area. The survey used a 10-call design witb one refusal conversion attempt on all soft refusals—i.e., refusals wben an excuse was offered or tbere was anotber reason wby calling back on anotber occasion migbt bave resulted in a completed survey. Refusal conversion letters were mailed to all listed
bousebolds wbo refused to participate in tbe survey. Reverse address lookup ' was also used based on RDD pbone numbers to provide furtber information on contacted bousebolds. Interviews were conducted in botb Spanisb and Englisb according to the preferences of tbe respondents, and compensation cbecks of $5.00 were mailed to tbeir bomes as tokens of appreciation for tbeir participation. All interviews were tape recorded; tbe average time for survey administration was 19 minutes. Tbe final sample size available for analysis in tbis article was 646 Harlem adults 18 years and older. Measures Survey measures were adapted for use among Harlem adults from previously fielded questionnaires on tecbnology use and bealtb information seeking, including (1) New York City Community Healtb Survey,^** (2) Daily Internet Tracking Survey," (3) U.S. Survey of Healtb and tbe Internet,^^ (4) Healtbstyles Survey,^» (5) Healtb Information National Trends Survey,-^" (6) Current Population Survey; Scbool Enrollment/Computer Use Supplement,^' and (7) New York City Social Indicators Survey.*^ All measures were self-reported and included tbe availability of computers and Internet access; future intentions for computer and Internet use; attitudes toward tbe Internet; frequency of Internet use in general and bealtb information seeking in particular; and perceived clarity, understanding, and trust of information found. Validity and reliability testing involved calling 15 volunteers recruited from previous focus groups. Seven of tbese 15 individuals agreed to participate in a second round of calls 2 weeks after tbe first interviews to determine test-retest reliability. Six of tbe 15 volunteers agreed to participate in postsurvey interviews to discuss tbe validity of selected survey questions. In terms of test-retest reliability, continuous variables bad a mean Pearson's correlation coefficient between tbe test and retest of .72. Discrete variables bad a mean percent agreement between the test and retest of .93. In terms of validity, three questions were identified as invalid, meaning that the understanding of
May/June 2011, Vol. 25, No. 5
Table 1 Sociodemographic Characteristics of Adults >18 Years in the HPSS, Harlem, and the United States Overall % of HPSS (N = 646)
% of Harlem
% of United States
Gender Female Male
67.8 32.2
62.7* 37.3
50.7t 49.3
Age group, y 18-24 25-34 35^4 45-54 55-64 >65
9.0 22.2 26.2 15.9 11.2 15.6
6.8 32.5
24.4
13.2 17.7 19.5 19.1 14.0 16.5
Race/efhnicify Black, non-Hispanic Whife, non-Hispanic or ofher Hispanic
56.6 10.8 32.6
50.9 16.6 32.5
12.1:1: 75.4 12.5
18.0 28.4 22.8 30.8
24.3 25.9 19.6 30.1
18.0§ 33.9 39.5 8.6
29.8 16.6 14.3 15.9 23.4
38.511 14.8 12.6 12.6 21.6
13.9 11.4 11.2 14.8 48.7
Healfh insurance coverage
89.8
84.6*
82.211
Employmenf In labor force Employed Unemployed Not in labor force (homemaker/sfudenf/refired)
75.0 75.4 24.6 25.0
50.9# 82.2 17.8 49.1
66.0** 95.3 4.7 34.0
Self-reported healfh Excellenf Very good Good Fair or poor
17.2 25.7 31.9 25.1
14.1* 23.7 33.5 28.8
35.6tt 29.9 22.7 11.8
Sociodemographic Characteristic
36.3
Education Grades 0-11 High school graduafe/equivalenf Some college/fechnical school Bachelor's degree or higher Annual household income
$50,000
HPSS indicafes Harlem Phone Survey Sample. * Dafa from fhe New York Cify Deparfmenf of Healfh and Menfal Hygiene,^^ t US Census Bureau: American FacfFinder: New York: Age and Sex, Dafa Set: 2006 American Communify Survey. Available af: hffp://facffinder. census.gov/servlef/STTable?_bm=y&-geo_id=04000US36&-qr_name=ACS_2006_EST_G00_S0101&-ds_name=ACS_2006_EST_G00_. Accessed January 22, 2008. t US Census Bureau: Census 2000 Brief: Overview of Race and Hispanic Origin. Available af: hffp://www.census.gov/prod/2001pubs/cenbr01-l.pdf. Accessed January 28, 2009. § US Census Bureau: American FacfFinder: New York: Educafional Affainmenf, Dafa Sef: 2006 American Communify Survey. Available af: http:// facffinder.census.gov/servlef/STTable?_bm=y&-qr_name=ACS_2006_EST_G00_S1501&-geo_id=04000US36&-ds_name=ACS_2006_EST_G00_&redoLog=false. Accessed January 22, 2008. II New York Cify Deparfmenf of Cify Planning: New York Cify Census FacfFinder: 2000 Census Profiles for New York Clfy. Available af: hffp://gis.nyc. gov/dcp/pa/address.jsp. Accessed Sepfember 29, 2009. H Cenfers for Disease Confrol and Prevention: Nafional Cenferfor Healfh Sfafisfics: Trends in Healfh Insurance Coverage by Poverty Sfafus Among Persons Under 65 Years of Age: Unifed Sfafes, 1997-2002. Available af: hffp://www.cdc.gov/nchs/producfs/pubs/pubd/hesfafs/insurance.hfm. Accessed January 22, 2008. # New York Cify Employmenf Sfafus for fhe Populafion 16 Years and Over by Sex: New York Cify Communify Disfricfs, 2000. Available af: hffp://www. nyc.gov/hfml/dcp/pdf/census/sf3lfp301 .pdf. Accessed January 28, 2009. (Nofe: percenfage "nof in labor force" differs, as numbers and percenfages reporfed in New York Cify are for fhose >16 years, compared wifh fhose >18 years in fhe HPSS.) ** US Deparfmenf of Labor: Bureau of Labor Sfafisfics: The Employmenf Sifuafion: March 2006. Available af: hffp://www.bls.gov/news.release/ archives/empsif_04072006.pdf. Accessed January 27, 2009. t t US Census Bureau: Healfh Insurance Coverage in 2005. Available af: hffp://wvw.census.gov/populafion/pop-profile/dynamic/Healfhlnsurance.pdf. Accessed January 24, 2008.
American Journal of Healfh Promofion
May/June 2011, Vol. 25, No. 5
327
Table 2 Percent of Computer Use and Internet Access Among Adults >18 Years in the HPSS, Specified HPSS Subgroups, and U.S. Adults Overall or Specified U.S. Adult Subgroups
% of HPSS (N = 646) Ever used a computer? Ever used the Internet? Current Internet user Frequency of Internet use Every day A couple times a week About once a week Less often Never Computer in home? No computer in home One or more computers in home
77.1 64.7 51.6
% Of Specified HPSS Subgroup
% of U.S. Adults or Specified U.S. Adult Subgroup*
"Computer ever" respondents, 84.1 "Internet ever" respondents, 80.2 Current Internet users
Not available! Not availablet All respondents, 73 Internet use from home
28.3 14.2 3.1 5.6 48.8
55.3 27.8 6.0 10.9
31.6 68;4
Internet connection in home
57.1
High-speed connection
43.3
60* 32 8
Not availablet Computer(s) in home 84.6 Internet connection in home 79.1
Not availablet Internet use from home 68
HPSS indicates Harlem Phone Survey Sample. * Pew Internet and American Life Project; August 2006—Health; [Q6a] "Do you use the Internet, at least occasionally?" and [Q6b] "Do you send or receive e-mail, at least occasionally?" Available at; http;//www.pewinternet.org/Shared-Content/Data-Sets/2006/August-2006-Health aspx Accessed September 29, 2009. t Equivalent questions for the same time period were unavailable for some items. t Pew Internet & American Life Project; August 2006—Health; [Q14] "How often do you use the Internet from home?" Available at; http;//www. pewinternet.org/Shared-ContenVData-Sets/2006/August-2006-Health.aspx. Accessed September 29, 2009. (Note; Answer choices included "several times a day," "about once a day," "3-5 days a week," "1-2 days a week," "once every few weeks," or "less often." Percentages were condensed to match the closest equivalent in the Harlem Phone Survey.)
the questions by the research staff differed from that of the participants. These questions were subsequently revised and retested. After the survey instrument was finalized, data were coded in a computer-assisted telephone interview system. Analysis
Data analysis occurred throughout 2008. Data were checked for outliers, and distributional properties were examined for all variables. Simple frequencies were calculated for all categorical variables, and means and standard deviations were calculated for all continuously scaled variables of interest. Selected prevalence estimates from the Harlem Phone Survey Sample (HPSS) were compared with other Harlem community-wide and U.S. national estimates.'^'^^^ Because our intent was to describe patterns of computer use and Internet access in different populations and because of important differences in survey designs and sample sizes, tests of significance
328
American Journal of Health Promotion
were not performed. Instead, the magnitudes of the differences in the estimates were considered as potentially important in distinguishing patterns of computer use, Internet access, and online health information searching in the compared populations.'^ For demographic characteristics and comparisons of current Internet users versus non-current users, x^ analyses were employed for discrete variables and /-tests were used for the analysis of continuous measures. Type I error was set at 5%. Because of the exploratory and descriptive nature of this study, no adjustments were made for multiple comparisons. RESULTS Deseription of Study Participants
A total of 646 Harlem residents completed the RDD landline phone survey. Fifty-one percent of the sample (n = 329) lived in Central Harlem, 31% (n = 198) lived in East Harlem, and 18% (n = 119) lived in Hamilton
Heights. In an attempt to include more men and a representative age distribution among participants, callers were screened for gender and age after the first 2 weeks of the survey. Of the 8249 phone nvimbers used, 661 were outright refusals; 4579 did not meet eligibility criteria for place of residence, age, and gender representation; 1659 were to unknown households; and the rest were a combination of numbers that were never answered or no contact was made for other reasons. These call statistics correspond to standard American Association for Public Opinion Research response rates of 23% to 35% and cooperation rates of 51% to 57%, depending upon the formulas used.'''' Demographic characteristics of the HPSS, the Harlem population, and the U.S. population are provided in Table 1. In general, the demographic characteristics of the HPSS matched those of the Harlem population. On the other hand, both the HPSS and the Harlem population were com-
May/June 2011, Vol. 25, No. 5
Table 3 Adults >18 Years in the Harlem Phone Survey Sample Who Are Current Users Versus Non-Current Users of the Internet by Sociodemographic Characteristics
Sociodemographic Characteristic
% of Current Users (n = 333)
% of NonCurrent Users (n = 313)
Gender Male Female
51.4 51.6
48.6 48.4
Age group, y 18-24 25-34 35-44 45-54 55-64 >65
69.0 66.2 61.9 47.1 41.7 14.0
31.0 33.8 38.1 52.9 58.3 86.0
Language(s) spoken English or English + other Spanish or other only
57.0 17.6
43.0 82.4
53.0 40.1 74.6
47.0 59.9 25.4
13.0 47.0 79.4
87.0 53.0 20.6
66.8 29.7 32.5
33.2 70.3 67.5
Annual household income $50,000
33.3 71.4 86.6
66.7 28.6 13.4
Health insurance coverage Yes No
52.0 44.6
48.0 55.4
Self-reported health Excellent Very good or good Fair Poor
72.1 60.0 22.3 7.3
27.9 40.1 77.7 92.7
Do your close friends or family use the Internet? Yes No
58.0 12.3
42.0 87.7
How many people that you know use the Internet? Most/some 58.9 Very few/none 16.5
41.1 83.5
High-speed Internet connection at home? Yes No
19.3 70.8
African-American/black Hispanic/Latino White Educational attainment Less than high school High school diploma/equivalent Some college or more
American Journal of Health Promotion
Computer Use and Internet Access
As presented in Table 2, 77% of the HPSS reported having used computers and 65% reported having used the Internet. To determine current Internet use, respondents were asked, "Have you ever gone online to access the Internet or World Wide Web or to send and receive e-mail?" and "Did you ever at some point use the Internet or e-mail but have since stopped for some reason?" Those who responded that they had ever gone online and had not stopped using the Internet or e-mail were considered current Internet users. Using this definition, approximately one-half (52%) of the HPSS were current Internet users, with 83% of these respondents reporting Internet use daily (55%) or at least a couple times per week (28%). Overall, 68% of the HPSS lived in a household with one or more computer; 85% of these respondents also had an Internet connection at home, regardless of whether or not they went online themselves, and 79% of those with an Internet connection in the home had a high-speed connection. In contrast, about one-third of the HPSS reported that they had never used the Internet (100% - 65% = 35%), and nearly half reported that they did not live in a household with an Internet connection (100% - 57% = 43%).
